The Multi-site Evaluation
of SVORI: Summary and
Synthesis
THE MULTI-SITE EVALUATION OF THE SERIOUS AND VIOLENT OFFENDER REENTRY INITIATIVE
December 2009
Pamela K. Lattimore
RTI International
3040 East Cornwallis Road
Research Triangle Park, NC 27709
Phone: (919) 485-7759
Fax: (919) 541-5985
Lattimore@rti.org
Christy A. Visher
University of Delaware/The Urban Institute
Center for Drug and Alcohol Studies
77 East Main Street
Newark, DE 19716
Phone: (302) 831-6921
Fax: (302) 831-3307
Visher@udel.edu
This project was supported by Grant No. 2004-RE-CX-002 awarded by the National Institute of Justice, Office of Justice Programs, U.S.
Department of Justice. Points of view in this document are those of the authors and do not necessarily represent the official position or policies
of the U.S. Department of Justice.
Acknowledgments
The Multi-site Evaluation of the Serious and Violent Offender
Reentry Initiative (SVORI) was supported by grant number
2004-RE-CX-002 from the National Institute of Justice (U.S.
Department of Justice) and was conducted by RTI International
and the Urban Institute. Points of view are those of the authors
and do not necessarily represent those of the U.S. Department
of Justice.
Principal Investigators
Pamela K. Lattimore, RTI International
Christy A. Visher, University of Delaware and Urban Institute
Report Authors
Pamela K. Lattimore, RTI International
Christy A. Visher, University of Delaware and Urban Institute
Staff Contributors
Susan Brumbaugh, RTI International
Alexander Cowell, RTI International
Debbie Dawes, RTI International
Christine Lindquist, RTI International
Mark Pope, RTI International
John Roman, Urban Institute
Danielle M. Steffey, RTI International
Laura Winterfield, Urban Institute
We also acknowledge the contributions of the site liaisons from
RTI and the Urban Institute, who documented the
implementation of SVORI programming across the sites and
facilitated data collection for the impact study. In addition, we
are grateful for the hard work and dedication shown by our field
interviewers, supervisors, and data collection task leader
throughout the data collection period.
iii
RTI and the Urban Institute thank the SVORI program
directors, other program and research staff from the SVORI
sites, and staff at the facilities where interviews were
conducted. We greatly appreciate the assistance and support
received from these individuals.
Finally, RTI and the Urban Institute acknowledge the invaluable
assistance and direction provided by the members of our
external advisory group.
For more information about the SVORI Multi-site Evaluation,
please visit our Web site at http://www.svori-evaluation.org/.
iv
Abstract
Statement of Purpose
The Serious and Violent Offender Reentry Initiative (SVORI)
funded 69 agencies in 2003 to develop programs to improve
criminal justice, employment, education, health, and housing
outcomes for released prisoners. These programs were to
conduct assessments and provide participants programs and
services during and after incarceration. The SVORI Multi-site
Evaluation was funded by the National Institute of Justice to
examine the extent to which SVORI program participation
improved access to appropriate, comprehensive, integrated
services and resulted in better outcomes.
Research Subjects
This report presents findings for 2,391 participants in 12 adult
and 4 juvenile sites selected for the impact evaluation (1,697
adult males, 357 adult females, and 337 juvenile males). The
study participants had extensive criminal and substance use
histories, low levels of education and employment skills, and
high levels of need across a range of services (e.g., education,
driver’s license, substance abuse treatment, and job training).
Study Methods
The impact evaluation included interviews 30 days pre-release
and 3, 9, and 15 months post-release. Data from state
agencies and the National Criminal Information Center
documented post-release recidivism. Propensity score
techniques were used to improve the comparability between
the SVORI and non-SVORI groups. Weighted analyses
examined the treatment effects of SVORI program
participation.
v
Major Findings
The report provides evidence that SVORI program participation
increased receipt of services and programming for the adults.
Program participants were significantly more likely, for
example, to have reentry plans, although levels of provision for
most services fell short of 100% and declined substantially
after release. The juvenile males received higher levels of
service pre-release than the adults, but there were few
differences between the SVORI and non-SVORI groups.
The results suggest modest improvements in outcomes for the
adult SVORI participants and few differences between the
juvenile SVORI and non-SVORI participants. SVORI programs
appear to have reduced substance use rates among program
participants, although overall drug use increased over time for
all groups and exceeded 50% at 15 months post release. For
the adult men, there were no differences in arrest and
reincarceration rates at 24 months (about 70% and 40%,
respectively). Women SVORI participants were significantly less
likely to have an arrest and more likely to have been
reincarcerated.
Conclusions
Although adult SVORI programs were successful in increasing
the types and amounts of needs-related services provided
before and after release from prison, the proportion of
individuals who reported receiving services was smaller than
the proportion that reported need and, generally, was smaller
than the proportion that the SVORI program directors expected
to have received services. This finding is consistent with the
fact that SVORI programs were still developing and
implementing their programs and serves as a reminder that
starting complex programs may require sustained effort over
several years to achieve full implementation.
Service delivery declined after release; therefore, overall, the
programs were unable to sustain support of individuals during
the critical, high-risk period immediately after release. This
decline may be due to the programs’ difficulty identifying and
coordinating services for individuals released across wide
geographic areas and, again, suggests the need for sustained
effort to achieve full implementation.
vi
SVORI program participation resulted in modest improvements
in intermediate outcomes for adults at levels consistent with
findings from meta-analyses of single-program efforts (e.g.,
10% to 20%). If the underlying model that links services to
improved intermediate outcomes that in turn improve
recidivism is correct, the level of improvement in these
intermediate outcomes may have been insufficient to result in
observable reductions in recidivism.
Additional analyses are planned to determine whether there are
specific programs or subgroups associated with positive
outcomes and to examine the relationship between receipt of
specific services and outcomes.
vii
Contents
Section
Abstract
Executive Summary
Page
v ES-1 Introduction
1 Research on Prisoner Reentry ..................................... 3 The Serious and Violent Offender Reentry Initiative .......12 Multi-site Evaluation Impact Sites ...............................15 Methods
19 Implementation Data Collection ..................................19 Identification and Enlistment of Impact Sites ................20 Identification of Comparison Respondent
Populations ........................................................24 Impact Data Collection: Interviews and
Administrative Data .............................................25 Propensity Score Models ............................................27 Respondent Characteristics
31 Adult Male SVORI Evaluation Participants .....................31 Adult Female SVORI Evaluation Participants .................33 Juvenile Male SVORI Evaluation Participants .................35 Comparison of Adult Male, Adult Female, and
Juvenile Male SVORI EvaluatIon respondents ..........38 Service Needs and Service Receipt
45 Adult Male Service Needs and Receipt .........................47 Adult Female Service Needs and Receipt ......................52 Juvenile Male Service Needs and Receipt .....................57 Comparison of Adult Male, Adult Female, and
Juvenile Male SVORI EvaluatIon Participants ...........63 Outcomes
79 Adult Male Outcomes ................................................79 Adult Female Outcomes.............................................87 Juvenile Male Outcomes ............................................93 ix
Comparison of Adult Male, Adult Female and
Juvenile Male Outcomes .......................................98 Economic Evaluation
113 Methodology Overview ............................................ 113 Sample ................................................................. 115 Data ..................................................................... 116 Results ................................................................. 117 Conclusions ........................................................... 119 x
Conclusions, Policy Implications, and Future Work
121 References
130 Appendix A. Data Tables
A-1 Exhibits
Exhibit Number
Page
ES-1. SVORI program logic model and evaluation
framework
ES-2 ES-2. Completed interviews by wave, SVORI status, and
demographic group
ES-7 1.
SVORI program logic model and evaluation
framework
13 2.
Program sizes among impact and non–impact
sites (as reported by program directors)
16
3.
Implementation status among adult impact and
non–impact sites
18 4.
Impact evaluation site selection criteria
22 5.
Original programs selected for the impact
evaluation
23 6.
Completed interviews by wave, SVORI status, and
demographic group
27 7.
Average ages at first arrest, instant incarceration,
and release
39 8.
Age at first use of alcohol, marijuana, and cocaine
40 9.
Percentages of respondents reporting use of
alcohol, marijuana, and cocaine, by SVORI
program status and demographic group
40 Numbers of prior arrests and convictions, by
demographic group
41 Percentages reporting person/violent, drug, and
property offenses associated with instant
incarceration, by demographic group
42 12-Item Short-Form mental and physical health
scales, by demographic group
43 13.
Items included in service need and service receipt
bundles
46 14.
Adult male pre-release service need bundle score
means across service type, by group
48 10.
11.
12.
xi
15.
Adult male service need bundle scores for Waves
1 through 4
49 16.
Adult male service receipt bundle scores across
waves by group
51 17.
Adult female pre-release service need bundle
scores across service type, by group
53 18.
Adult female service need bundle scores for
Waves 1 through 4
55 19.
Adult female service receipt bundle scores across
waves, by group
57 20.
Juvenile male pre-release service need bundle
scores across service type, by group
58 21.
Juvenile male service need bundle scores for
Waves 1 through 4
60 22.
Juvenile male service receipt bundle scores across
waves, by group
62 23.
Weighted service need super-bundle scores by
demographic group, wave, and study group
64 24.
Employment/education/skills need bundle scores
by demographic group, study group, and data
collection wave
65 Transition services need bundle scores by
demographic group, study group, and data
collection wave
67 26.
Health needs bundle scores by demographic
group, study group, and data collection wave
68 27.
Service receipt super-bundle scores by
demographic group, study group, and data
collection wave
70 Coordination service receipt bundle scores by
demographic group, study group, and data
collection wave
71 Employment/education/skills service receipt
bundle scores by demographic group, study
group, and data collection wave
73 Transition receipt service bundle scores by
demographic group, study group, and data
collection wave
75 Health services receipt bundle scores by
demographic group, study group, and data
collection wave
77 Weighted means and regression parameter
estimates for housing outcomes (adult males)
80 Weighted means and regression parameter
estimates for employment outcomes (adult males)
81 25.
28.
29.
30.
31.
32.
33.
xii
34.
Weighted means and regression parameter
estimates for substance use outcomes (adult
males)
83 Weighted means and regression parameter
estimates for core self-reported recidivism
outcomes (adult males)
84 34.
Official measures of recidivism (adult males)
86 37.
Weighted means and regression parameter
estimates for housing outcomes (adult females)
87 38.
Weighted means and regression parameter
estimates for employment outcomes (adult
females)
88 Weighted means and regression parameter
estimates for substance use outcomes (adult
females)
90 Weighted means and regression parameter
estimates for core self-report recidivism outcomes
(adult females)
91 41.
Official measures of recidivism (adult females)
92 42.
Weighted means and regression parameter
estimates for housing outcomes (juvenile males)
94 43.
Weighted means and regression parameter
estimates for employment outcomes (juvenile
males)
95 Weighted means and regression parameter
estimates for substance use outcomes (juvenile
males)
96 Weighted means and regression parameter
estimates for core self-report recidivism outcomes
(juvenile males)
97 46.
Housing independence by demographic group,
study group, and data collection wave
99 47.
Support of self with job by demographic group,
study group, and data collection wave
101 48.
Job with formal pay by demographic group, study
group, and data collection wave
102 49.
Job with benefits by demographic group, study
group, and data collection wave
104 No self-reported drug use in the past 30 days or
positive drug test by demographic group, study
group, and data collection wave
105 35.
39.
40.
44.
45.
50.
51 Differences in drug use in the past 30 days by
demographic group, study group, and data
collection wave
52.
Self-reported compliance with supervision
conditions by demographic group, study group,
and data collection wave
106 108 xiii
53.
xiv
Self-reported noncommission of criminal activity
by demographic group, study group, and data
collection wave
109 54.
Cumulative arrest rates for adult SVORI and nonSVORI study participants
110 55.
Cumulative reincarceration rates for adult SVORI
and non-SVORI study participants
111 A-1.
Completed interviews by wave, by demographic
group and site
A-1 A-2.
Weighted means and parameter estimates of the
effect of SVORI on service need (adult males)
A-2 A-3.
Weighted means and parameter estimates of the
effect of SVORI on service receipt (adult males)
A-3 A-4.
Service needs weighted means and SVORI
parameter estimates (adult females)
A-5 A-5.
Weighted means and parameter estimates of the
effect of SVORI on service receipt
A-6 A-6.
Service needs weighted means and SVORI
parameter estimates (juvenile males)
A-8 A-7.
Service receipt weighted means and SVORI
parameter (juvenile males)
A-10 Executive Summary
SVORI responded to
emerging research
findings that suggested
providing individuals
with comprehensive,
coordinated services
based on needs and risk
assessments can result in
improved post-release
outcomes.
In 2003 the U.S. Department of Justice, Department of Labor,
Department of Education, Department of Housing and Urban
Development, and Department of Health and Human Services
provided more than $100 million in grant funds to states to
develop, enhance, or expand programs to facilitate the reentry
of adult and juvenile offenders to communities from prisons or
juvenile detention facilities. The Serious and Violent Offender
Reentry Initiative (SVORI) funded agencies to develop
programs to improve criminal justice, employment, education,
health, and housing outcomes for released prisoners. Sixty-nine
agencies received federal funds ($500 thousand to $2 million
over 3 years) to develop 89 programs.
The initiative responded to emerging research findings that
suggested that providing individuals with comprehensive,
coordinated services based on needs and risk assessments can
result in improved post-release outcomes. Grantees were to
use their SVORI funding to create a three-phase continuum of
services for returning serious or violent prisoners that began
during the period of incarceration, intensified just before
release and during the early months post-release, and
continued for several years after release as former inmates
took on more productive and independent roles in the
community.
The SVORI logic model identifies SVORI funding, technical
assistance, and requirements as inputs that, in combination
with local resources in the sites (throughputs), yield a set of
services and programming (outputs) expected to improve the
intermediate and recidivism outcomes for SVORI participants,
as well as improve the state and local systems that provide the
services and programs (Exhibit ES-1). Community and
individual participant characteristics influence these
ES-1
Multi-site Evaluation of SVORI: Summary & Synthesis
Exhibit ES-1. SVORI program logic model and evaluation framework
Community Context
Offender Context
Population Characteristics
Criminal History
Mental & Physical Health
Substance Abuse
Education/Training/Work Experience
Family Ties
Population Characteristics
Unemployment Rates
Service Availability
Residential Stability
Post-release Supervision Structure
Inputs:
The SVORI
Federal Funding &
Other Resources
Technical Assistance
Federal Grant
Requirements
Throughputs
Local Partnership
Formation &
Functioning
State & Local
Resources
Outputs:
Implementation
In-Prison
Coordination/Supervision
Education/Training
Family Services
Health Services
Transition Services
Community
Coordination/Supervision
Education/Training
Family Services
Health Services
Transition Services
Post-Supervision
Outcomes
Offender
Community Involvement
Employment
Family Contact/Stability
Health/Mental Health
Housing
Recidivism
Substance Use
Supervision Compliance
Systems
Rearrest Rates
Reincarceration Rates
Systems Change
Community Reintegration
Activities
Implementation Assessment
Evaluation
Components
Impact
Evaluation
Cost-Benefit Analysis
throughputs, outputs, and outcomes. The model shows that
SVORI was an outcome- or goal-oriented initiative that
specified outcomes, or goals, that should be achieved by
programs developed locally. Criteria specified by the federal
partners for the local programs were the following:1
1
ES-2
ƒ
Programs were to improve criminal justice, employment,
education, health (including substance use and mental
health), and housing outcomes.
ƒ
Programs were to include collaborative partnerships
between correctional agencies, supervision agencies,
other state and local agencies, and community and
faith-based organizations.
In some cases, grantees asked for and received exceptions to these
criteria. For example, some programs were primarily post-release
programs, and age restrictions were sometimes lifted (e.g., for
programs targeting sex offenders).
Executive Summary
SVORI was an outcomeor goal-oriented initiative
that specified outcomes,
or goals, that should be
achieved by programs
that were developed
locally.
ƒ
Program participants were to be serious or violent
offenders.
ƒ
Program participants were to be 35 years of age or
younger.
ƒ
Programs were to encompass three stages of reentry—in
prison, post-release on supervision, and postsupervision.
ƒ
Needs and risk assessments were to guide the provision
of services and programs to participants.
The SVORI programs attempted to address the initiative’s goals
and provide a wide range of coordinated services to returning
prisoners. Although SVORI programs shared the common goals
of improving outcomes across various dimensions and
improving service coordination and systems collaboration,
programs differed substantially in their approaches and
implementations (Lindquist, 2005; Winterfield & Brumbaugh,
2005; Winterfield, Lattimore, Steffey, Brumbaugh, & Lindquist,
2006; Winterfield & Lindquist, 2005).
In spring 2003, the National Institute of Justice awarded RTI
International, a nonprofit research organization, a grant to
evaluate programs funded by SVORI. The Urban Institute, a
nonpartisan economic and social policy research organization,
collaborated on the project. With data collected from grantee
staff, partnering agencies, and returning prisoners, the 6-year
evaluation involved an implementation evaluation of all 89
SVORI programs, an intensive impact evaluation of 12 adult
and 4 juvenile programs, and an economic analysis of a subset
of the impact sites (see Lattimore, Visher, Winterfield,
Lindquist, & Brumbaugh, 2005). The goal of the SVORI
evaluation was to document the implementation of SVORI
programs and determine whether they accomplished SVORI’s
overall goal of increasing public safety by reducing recidivism
among the populations served. The SVORI evaluation was
designed to answer the following research questions:
ƒ
To what extent did SVORI lead to more coordinated and
integrated services among partner agencies?
ƒ
To what extent did SVORI participants receive more
individualized and comprehensive services than
comparable, non-SVORI offenders?
ƒ
To what extent did reentry participants demonstrate
better recidivism, employment, health, and personal
ES-3
Multi-site Evaluation of SVORI: Summary & Synthesis
functioning outcomes than comparable, non-SVORI
offenders?
ƒ
The evaluation was
designed to determine
whether individuals who
participated in enhanced
reentry programming, as
measured by their
enrollment in SVORI
programs, had improved
post-release outcomes.
To what extent did the benefits derived from SVORI
programming exceed the costs?
The local nature of the SVORI programs and the expectation
that programs would tailor services to meet individual needs
meant that the intervention to be evaluated was not a program
in the typical conceptualization of the term (e.g., a residential
drug program or a cognitive behavior program). Instead,
SVORI was a funding stream that agencies used to expand and
enhance existing programs or to develop and implement new
programs. Further, individuals not in SVORI programs also
generally received some services. Thus, although the
components of the individual programs were identified and the
extent of service receipt was measured, the SVORI Multi-site
Evaluation was not designed to examine the impact of specific
services or combinations of services. The evaluation was
designed to determine whether individuals who participated in
enhanced reentry programming, as measured by their
enrollment in SVORI programs, had improved post-release
outcomes.
This report summarizes and synthesizes findings for the three
demographic groups included in the evaluation—adult males,
adult females, and juvenile males. Results from the impact and
economic evaluations are presented in the following reports:
ES-4
ƒ
Lindquist, C. H., Barrick, K., Lattimore, P. K., & Visher,
C. A. (2009). Prisoner reentry experiences of adult
females: Characteristics, service receipt, and outcomes
of participants in the SVORI Multi-site Evaluation.
Research Triangle Park, NC: RTI International.
ƒ
Lattimore, P. K., Steffey, D. M., & Visher, C. A. (2009).
Prisoner reentry experiences of adult males:
Characteristics, service receipt, and outcomes of
participants in the SVORI Multi-site Evaluation. Research
Triangle Park, NC: RTI International.
ƒ
Hawkins, S., Dawes, D., Lattimore, P. K., & Visher, C. A.
(2009). Reentry experiences of confined juvenile
offenders: Characteristics, service receipt, and outcomes
of juvenile male participants in the SVORI Multi-site
Evaluation. Research Triangle Park, NC: RTI
International.
ƒ
Cowell, A., Roman, J., & Lattimore, P. K. (2009). An
economic evaluation of the Serious and Violent Offender
Reentry Initiative. Research Triangle Park, NC: RTI
International.
Executive Summary
The methods incorporated in the evaluation are described in the
following report:
ƒ
Lattimore, P. K., & Steffey, D. M. (2009). The Multi-Site
Evaluation of SVORI: Methodology and analytic
approach. Research Triangle Park, NC: RTI
International.
The following sections summarize prior reentry research and
the research design. Subsequent sections present key findings
and conclusions.
RESEARCH ON PRISONER REENTRY
In 2008, more than 735,000 prisoners were released from state
and federal prisons across the country (West, Sabol, & Cooper,
2009). This number represents a greater than four-fold
increase over the nearly 170,000 released in 1980 (Harrison,
2000). With the exception of those who die while in prison, all
prisoners will eventually “reenter” the community. Nationwide,
more than half of individuals who are released from prison are
reincarcerated within 3 years. Programs and services for men
and women leaving prison are designed to stop this “revolving
door” and encourage individuals to desist from offending.
Until recently, the majority of rehabilitation and reentry
strategies have been dominated by service providers who
represent a single domain from among the possible correlates
of desistance. For instance, many reentry programs are
concentrated in one-stop workforce centers whose main
function is to prepare and place individuals in jobs. Reentry
services may include interventions directly related to skill
acquisition to improve labor market prospects, interventions
such as job readiness, training, and placement programs. Other
reentry programs may focus on reducing specific deficits by
reducing substance abuse, addressing physical and mental
health disorders, improving educational attainment through
General Education Development (GED) or high school
programming, or offering other assistance from the small
(access to official identification and transfer of prescriptions) to
the large (securing transitional and long-term housing).
Reentry initiatives also may assist in the cognitive development
of participants to promote behavioral change through faithbased or classroom-based programming (e.g., anger
management, parenting skills, life skills).
ES-5
Multi-site Evaluation of SVORI: Summary & Synthesis
The complexity of the
disadvantages
confronting prisoners
after release means that
individual offenders often
require more than a
single program or
intervention.
Many reentry specialists
are encouraging a
broader focus on
comprehensive reentry
strategies, not specific
programs.
However, the needs of individuals returning to the community
usually span these domains of problems, and typical service
providers are unlikely to be as effective at providing or
facilitating other services as they are in their primary area of
expertise. For example, it is not unusual for individuals
struggling with mental health and substance abuse disorders to
be denied entry into programs designed to respond to either
but not both of these disorders. The complexity of the
disadvantages confronting prisoners after release means that
individual offenders often require more than a single program
or intervention. To address this dilemma, many reentry
specialists are encouraging a broader focus on comprehensive
reentry strategies, not specific programs (Lattimore, 2007;
National Research Council, 2007; Reentry Policy Council, 2005;
Visher, 2007). Such strategies would involve multiple levels of
government, coordination of efforts across agencies, and
involvement of organizations traditionally not part of the
reentry discussion (e.g., public health, local businesses,
community colleges). Moreover, these coordinated efforts
would perhaps improve reintegration across a range of
outcomes (e.g., employment, substance use, health) broader
than reductions in recidivism.
This emerging focus on the need for comprehensive
programming provided the context within which the federal
government developed the Serious and Violent Offender
Reentry Initiative, resulting in the award of SVORI grants in
2002 and SVORI program start-ups in 2003 and 2004. This
brief review of previous research provides a context in which
the findings of the Multi-Site Evaluation of the Serious and
Violent Offender Reentry Initiative can be assessed.
RESEARCH DESIGN
The SVORI Multi-site Evaluation included an implementation
assessment (to document the programming delivered across
the SVORI programs) and an impact evaluation (to determine
the effectiveness of programming). An economic analysis was
also conducted in five of the impact sites to assess the extent
to which program benefits exceeded costs; findings from this
study are reported separately (see Cowell et al., 2009).
Sixteen programs were included in the impact evaluation,
comprising 12 adult programs and 4 juvenile programs located
ES-6
Executive Summary
in 14 states (adult only unless specified): Colorado (juveniles
only), Florida (juveniles only), Indiana, Iowa, Kansas (adults
and juveniles), Maine, Maryland, Missouri, Nevada, Ohio,
Oklahoma, Pennsylvania, South Carolina (adults and juveniles),
and Washington. The impact evaluation included pre-release
interviews (conducted approximately 30 days before release
from prison) and a series of follow-up interviews (conducted at
3, 9, and 15 months post-release). In addition, oral swab drug
tests were conducted during the 3- and 15-month interviews
for respondents who were interviewed in a community setting.
Recidivism data were obtained from the Federal Bureau of
Investigation National Crime Information Center (NCIC) and
state correctional and juvenile justice agencies. Nearly 2,400
prisoners returning to society—some of whom received SVORI
programming and some of whom received “treatment as usual”
in their respective states—were included in the impact
evaluation.
The numbers of completed interviews by SVORI status and
demographic group are shown in Exhibit ES-2. For the
distributions of these cases by site, see Appendix Exhibit A-1.
Exhibit ES-2. Completed interviews by wave, SVORI status, and demographic group
State
Adult males
Adult females
Juvenile males
Total
Wave 1
NonSVORI SVORI
863
834
153
204
152
185
1,168
1,223
Wave 2
NonSVORI SVORI
529
455
110
134
105
131
744
720
Wave 3
NonSVORI SVORI
565
470
119
134
108
131
792
735
Wave 4
NonSVORI SVORI
582
531
124
152
107
141
813
824
Note: Wave 1 = 30 days pre-release; Wave 2 = 3 months post-release; Wave 3 = 9 month post-release; Wave 4 = 15
months post-release.
Although the response rates were reasonable, the possibility
remains that respondents who “dropped out” of subsequent
waves of interviews differed from those who completed the
follow-up interviews. As preliminary evidence that the attrition
was random or affected the SVORI and non-SVORI groups
similarly, the SVORI and comparison groups were compared
and were found to be similar at each wave on a range of
characteristics. Results from models that examined for
differences between groups with respect to response also
ES-7
Multi-site Evaluation of SVORI: Summary & Synthesis
suggested that SVORI program participation was not related to
whether a participant responded.
Propensity score techniques were used to improve the
comparability between the SVORI and non-SVORI groups.
Weighted analyses were used to examine the treatment effects
of SVORI program participation with respect to outcomes in
housing, employment, family/peer/community involvement,
substance use, physical and mental health, and criminal
behavior and recidivism.
KEY FINDINGS
This section summarizes key findings from the evaluation.
Characteristics of study participants are described next,
followed by descriptions of expressed service need, reported
service receipt, and post-release outcomes.
Research Subject Characteristics
The study participants
were high-risk offenders
who had extensive
criminal and substance
use histories, low levels
of education and
employment skills, and
families and peers who
were substance and
criminal justice system
involved.
The study participants were high-risk offenders who had
extensive criminal and substance use histories, low levels of
education and employment skills, and families and peers who
were substance and criminal justice system involved. There
were few statistically significant differences in the
characteristics of the groups.
Nearly all of the respondents reported having used alcohol and
drugs during their lifetimes. Most reported having used one or
more illicit drugs during the 30 days before their current
incarceration.
SVORI and non-SVORI respondents reported considerable
involvement with the criminal justice system before their
current incarceration. On average, the adult male respondents
were 16 years old at the time of their first arrest and had been
arrested more than 12 times.2 The average adult female
respondent was 19 years old at the time of her first arrest and
had been arrested more than 10 times. The average juvenile
male respondent was 13 at the time of his first arrest and had
been arrested more than 6 times. In addition to their current
terms of incarceration, most adult respondents had served a
previous prison term, and most of the juvenile males reported
multiple detentions.
2
ES-8
This measure of prior arrest recoded extreme values to the 95th
percentile of reported arrests.
Executive Summary
Offenders returning to
their communities after
serving time comprise a
population with extremely
high needs.
The findings substantiate previous research that offenders
returning to their communities after serving time in prison (or
juvenile detention) comprise a population with extremely high
needs. The expressed needs remained high (if somewhat
diminished from pre-release) up to 15 months after release
from prison. Overall, there was little difference in reported
needs between the SVORI and non-SVORI groups.
The report provides evidence that adults participating in SVORI
programs received more services and programming, including
programs to prepare for release, meeting with a case manager,
and receiving a needs assessment—although levels of provision
for most services fell far short of 100%, were substantially
below expressed needs for services, and declined substantially
after release. Although juvenile subjects received higher levels
of services than the adults, on average, there were few
differences between SVORI and non-SVORI groups. Service
receipt for both SVORI and non-SVORI respondents was highest
during confinement.
Post-Release Outcomes
For the adults, the significant—albeit less-than-universal—
increase in service receipt associated with participation in
SVORI programs was associated with moderately better
outcomes with respect to housing, employment, substance use,
and self-reported criminal behavior, although these
improvements were not associated with reductions in official
measures of reincarceration. As many of the previous
evaluations of reentry programs have focused primarily on
recidivism and substance use, this evaluation provided an
opportunity to examine the impacts of reentry programming on
an array of other important indicators of successful
reintegration, including housing and employment.
The two groups of juvenile males showed few differences in
outcomes. SVORI respondents were significantly more likely
than non-SVORI respondents to be in school 3 months after
release from confinement. Fifteen months after release, SVORI
respondents were much more likely to have a job with benefits.
No significant differences were found between SVORI and nonSVORI respondents in substance use, physical health, mental
health, or recidivism outcomes.
ES-9
Multi-site Evaluation of SVORI: Summary & Synthesis
Economic Evaluation
The economic evaluation was conducted in four adult sites
(males only) and one juvenile site and included an analysis of
pre-release costs and a cost-benefit analysis. Separate analyses
were conducted for the adult and juvenile sites. The evaluation
showed considerable variability among the sites in pre-release
costs, with the incremental costs of SVORI pre-release
programming ranging from $658 to $3,480 in average
additional costs. The domain that contributed the most to the
difference in pre-release costs was case management, although
in one site the difference between SVORI and non-SVORI
service receipt was driven by employment/education/life skills
and substance use services.
The cost-benefit analysis combined service costs with postrelease criminal justice costs to generate estimates of net
costs. Criminal justice costs were based on official arrest and
reincarceration findings. The estimates of net costs had
substantial variability and did not show statistically significant
differences between SVORI and non-SVORI groups.
Individuals interested in the methodology and detailed findings
from the economic evaluation are referred to Cowell et al.
(2009).
CONCLUSIONS
Adult SVORI programs were successful in increasing the types
and amounts of needs-related services provided to participants
before and after release from prison; however, the proportion
of individuals who reported having received services was
smaller than the proportion that reported need and, generally,
was smaller than the proportion that the SVORI program
directors expected to have received services. This finding is
consistent with the fact that SVORI programs were still
developing and implementing their programs and serves as a
reminder that starting complex programs may require
sustained effort over several years to achieve full
implementation.
Service delivery declined after release; therefore, overall, the
programs were unable to sustain support of individuals during
the critical, high-risk period immediately after release. This
decline may be due to the programs’ difficulty identifying and
coordinating services for individuals released across wide
ES-10
Executive Summary
geographic areas and, again, suggests the need for sustained
effort to achieve full implementation.
SVORI program participation resulted in modest improvements
in intermediate outcomes for adults at levels consistent with
findings from meta-analyses of single-program efforts (e.g.,
10% to 20%). If the underlying model that links services to
improved intermediate outcomes that in turn improve
recidivism is correct, the level of improvement in these
intermediate outcomes may have been insufficient to result in
observable reductions in recidivism.
Results from the evaluation of the four programs for juvenile
males showed fewer differences than the adults did between
the two groups in services provided. SVORI program
participants were significantly more likely than non-SVORI
respondents to report having received employment-related
services before release and in the first 3 months after release.
As with the adults, the needs expressed by the juvenile males
were higher—sometimes substantially higher—than reported
receipt of services and programming. The findings suggest that
programs should apply additional effort in evaluating the levels
of services adequate to meet the expressed needs of these
young serious offenders.
Additional analyses are planned to determine whether specific
programs or subgroups are associated with positive outcomes
and to examine the relationships between receipt of specific
services and outcomes.
ES-11
Introduction
The Serious and Violent Offender Reentry Initiative (SVORI)
was a collaborative federal effort to improve outcomes for
adults and juveniles returning to their communities after a
period of incarceration. The initiative sought to help states
better use their correctional resources to address outcomes
along criminal justice, employment, education, health, and
housing dimensions by providing more than $100 million in
grant funds in 2003 to state agencies to establish or enhance
prisoner reentry programming. Funded by the U.S.
Departments of Justice, Labor, Education, Housing and Urban
Development, and Health and Human Services, SVORI was an
unprecedented national response to the challenges of prisoner
reentry. Sixty-nine agencies received federal funds ($500
thousand to $2 million over 3 years) to develop 89 programs.
Across the grantees, programming was provided to adult
males, adult females, and juveniles.
The initiative responded to emerging research findings that
suggested that providing individuals with comprehensive,
coordinated services based on needs and risk assessments can
result in improved post-release outcomes. SVORI funding was
intended to create a three-phase continuum of services for
returning prisoners that began during the period of
incarceration, intensified just before release and during the
early months post-release, and continued for several years
after release as former inmates took on more productive and
independent roles in the community. The SVORI programs
attempted to address the initiative’s goals and provide a wide
range of well-coordinated services to returning prisoners.
Although SVORI programs shared the common goals of
improving outcomes across various dimensions and improving
service coordination and systems collaboration, programs
1
Multi-site Evaluation of SVORI: Summary & Synthesis
differed substantially in their approaches and implementations
(Lindquist, 2005; Winterfield & Brumbaugh, 2005; Winterfield
et al., 2006; Winterfield & Lindquist, 2005).
The SVORI Multi-site Evaluation was funded by the National
Institute of Justice in the spring of 2003. This evaluation
included an implementation assessment (to document the
programming delivered across the SVORI programs), an impact
evaluation (to determine the effectiveness of programming),
and a cost-benefit evaluation (to determine whether program
benefits exceeded program costs. All 89 programs were
included in the implementation evaluation. Twelve adult and
four juvenile programs in 14 states were included the impact
evaluation. A subset of the impact sites (four adult and one
juvenile program) was included in the cost-benefit evaluation.
This Summary and Synthesis combines the SVORI Multi-site
Evaluation findings presented in the following reports:
ƒ
Lattimore, P. K., & Steffey, D. M. (2009). The Multi-Site
Evaluation of SVORI: Methodology and analytic
approach. Research Triangle Park, NC: RTI
International.
ƒ
Lindquist, C. H., Barrick, K., Lattimore, P. K., & Visher,
C. A. (2009). Prisoner reentry experiences of adult
females: Characteristics, service receipt, and outcomes
of participants in the SVORI Multi-site Evaluation.
Research Triangle Park, NC: RTI International.
ƒ
Lattimore, P. K., Steffey, D. M., & Visher, C. A. (2009).
Prisoner reentry experiences of adult males:
Characteristics, service receipt, and outcomes of
participants in the SVORI Multi-site Evaluation. Research
Triangle Park, NC: RTI International.
ƒ
Hawkins, S., Dawes, D., Lattimore, P. K., & Visher, C. A.
(2009). Reentry experiences of confined juvenile
offenders: Characteristics, service receipt, and outcomes
of juvenile male participants in the SVORI Multi-site
Evaluation. Research Triangle Park, NC: RTI
International.
Findings from the economic study in five impact sites are
presented in the following report:
ƒ
2
Cowell, A., Roman, J., & Lattimore, P. K. (2009). An
economic evaluation of the Serious and Violent Offender
Reentry Initiative. Research Triangle Park, NC: RTI
International.
Introduction
In the remainder of this chapter, previous research on prisoner
reentry is summarized, the SVORI and the evaluation design
are described, and the SVORI programs provided in the 12
impact evaluation sites are characterized. Subsequent chapters
provide detailed information on characteristics of the evaluation
participants, self-reported service needs and receipt at each of
the four interviews, post-release outcomes, and conclusions
and policy recommendations.
RESEARCH ON PRISONER REENTRY
In 2008, more than 735,000 prisoners were released from state
and federal prisons across the country (West et al., 2009). This
number represents a greater than four-fold increase over the
nearly 170,000 released in 1980 (Harrison, 2000). Juveniles
also play a substantial role in crime in the United States. In
2007 an estimated 2.18 million youth were arrested
(Puzzanchera, 2009). Data from the 2006 Juvenile Residential
Facility Census show that approximately 95 thousand juveniles
were held in juvenile facilities (Sickmund, Sladky, & Kang,
2008), and among this total about 65 thousand were
committed, meaning they were placed in the facility by a courtordered disposition.
From a developmental perspective, juvenile confinement often
leads to inadequate preparation for young adulthood, and a
juvenile’s delinquent involvement is likely to manifest in adult
criminality (McCord, 1992). Snyder and Sickmund (2006)
report that approximately “one quarter of juveniles who
offended at ages 16–17 also offended as adults at ages 18–19.”
With the exception of those who die while in prison, all
prisoners will eventually “reenter” the community.
This section summarizes research on the issues associated with
and the programmatic responses to adult reentry and then
briefly summarizes additional considerations for juveniles
returning to their communities from detention.
Prisoner Reentry
Prisoner reentry has sweeping consequences for the individual
prisoners themselves, their families, and the communities to
which they return (Petersilia, 2003; Travis, 2005). Nationwide,
over half of individuals who are released from prison are
reincarcerated within three years. Programs and services for
3
Multi-site Evaluation of SVORI: Summary & Synthesis
men and women leaving prison are designed to stop this
revolving door and encourage individuals to desist from
offending. Imprisonment without such preparation for
community reintegration may reduce human capital and impede
the acquisition of pro-social skills and behaviors, thus lessening
the probability of a successful transition from prison to the
community (Visher & Travis, 2003; Western, 2007). However,
in comparison to twenty years ago, men and women leaving
prison are less prepared for reintegration, less connected to
community-based social structures, and more likely to have
health or substance abuse problems than prior cohorts (Lynch
& Sabol, 2001; Petersilia, 2005).
In recent years, significant attention has been focused on the
impact of these increases in rates of incarceration and rates of
return from jail or prison (Bonczar & Beck, 1997; Clear, Rose, &
Ryder, 2001; Hagan & Coleman, 2001; Mauer, 2000; Travis,
2005). The geographic clustering of former prisoners by socioeconomic characteristics has led to disproportionate rates of
removal from, and return to, already distressed communities
(Clear et al., 2001; Lynch & Sabol, 2001). As a result, current
research on the social and economic impacts of incarceration is
increasingly focused on local effects of incarceration and
prisoner reintegration, and the concurrent effects on family
structure, intergenerational offending, and general community
well-being (Clear et al., 2001; Hagan & Dinovitzer, 1999).
Prisoner reentry programs that have emerged since the late
1990s seek to address the effects of incarceration by more
successfully reintegrating former prisoners, thereby reducing
subsequent offending.
Reentry programming is designed to break the cycle between
offending and incarceration. Incarcerating offenders generally
has two purposes: incapacitation and deterrence. Incapacitation
leads to temporary instrumental desistance, and specific
deterrence may lead to future deterrence. However, desistance
is mainly achieved through rehabilitative programming.
Predictors of desistance generally do not vary by the pattern of
past criminal behavior or by the antecedent characteristics of
the offender (Laub & Sampson, 2001). Processes that
consistently are identified as leading to desistance include
marriage and stable families (Laub & Sampson, 1993; Sampson
& Laub, 1993), aging (Glueck & Glueck, 1974; Gottfredson &
Hirschi, 1990; Laub & Sampson, 2003), stable employment
4
Introduction
(Laub & Sampson, 1993; Sampson & Laub, 1993) and reduced
exposure to antisocial peers (Laub & Sampson, 2003; Warr,
1998). In addition, all of these outcomes may be dependent
upon cognitive changes in identity which are the precursor to
changes in behavior (Maruna, 2001).
Until recently, the majority of rehabilitation and reentry
strategies have been dominated by service providers who
represent a single domain from among the possible correlates
of desistance. For instance, many reentry programs are
centered on one-stop workforce centers whose main function is
to prepare and place individuals in jobs. Reentry services may
include interventions directly related to skill acquisition to
improve labor market prospects such as job readiness, training,
and placement programs. Other reentry programs may focus on
reducing specific deficits by reducing substance abuse,
addressing physical and mental health disorders, improving
educational attainment through GED or high school
programming, or offering other assistance from the small
(access to official identification and transfer of prescriptions) to
the large (securing transitional and long-term housing).
Reentry initiatives also may assist in the cognitive development
of participants to promote behavioral change through faithbased or classroom-based programming (e.g., anger
management, parenting skills, life skills).
The complexity of the
disadvantages
confronting prisoners
after release means that
individual offenders often
require more than a
single program or
intervention.
Many reentry specialists
are encouraging a
broader focus on
comprehensive reentry
strategies, not specific
programs.
However, the needs of individuals returning to the community
usually span these domains of problems, and typical service
providers are unlikely to be as effective at providing or
facilitating other services as they are in their primary area of
expertise. For example, it is not unusual for individuals
struggling with mental health and substance abuse disorders to
be denied entry into programs designed to respond to either
but not both of these disorders. The complexity of the
disadvantages confronting prisoners after release means that
individual offenders often require more than a single program
or intervention. To address this dilemma, many reentry
specialists are encouraging a broader focus on comprehensive
reentry strategies, not specific programs (Lattimore, 2007;
National Research Council, 2007; Re-entry Policy Council,
2005; Visher, 2007). Such strategies would involve multiple
levels of government, coordination of efforts across agencies,
and involvement of organizations that are traditionally not part
of the reentry discussion (e.g., public health, local businesses;
5
Multi-site Evaluation of SVORI: Summary & Synthesis
community colleges). Moreover, these coordinated efforts may
improve reintegration across a broader range of outcomes
(e.g., employment, substance use, health) than simply
reductions in recidivism.
This emerging focus on the need for comprehensive
programming provided the context within which the federal
government developed the Serious and Violent Offender
Reentry Initiative, resulting in the award of SVORI grants and
SVORI program start-ups in 2003 and 2004. This brief review
of literature provides a context within which the findings of the
Multi-Site Evaluation of the Serious and Violent Offender
Reentry Initiative can be assessed.
Foremost among the
challenges is the lack of
theoretical models that
articulate behavior
change among former
prisoners.
Because reentry programs for individuals exiting prison are
relatively new, few impact evaluations of programs exist that
focus specifically on reentry (Petersilia, 2004). Additionally, the
extant research assessing the effectiveness of programs for
formerly incarcerated individuals, whether focused on reentry
or general rehabilitation, is burdened with substantial
challenges. Foremost among the challenges is the lack of
theoretical models that articulate behavior change among
former prisoners. Within any particular substantive area, there
are also problems of fidelity in that a particular service
approach may manifest itself in different ways under different
programs and circumstances. As a result, it is often difficult to
generalize research findings from one program to others, and
substantial variability exists among the outcome variables
examined (e.g., employment, homelessness, substance use,
recidivism). The numerous combinations of program types
unique to each study also render comparisons difficult. Finally,
there are problems related to the research itself, as rigorous
experimental designs—including the use of comparison groups
(randomly assigned or otherwise)—are rare in this research
literature (National Research Council, 2007).
Several reviews of reentry program evaluations recently have
examined the available research on what works with regard to
reentry and/or rehabilitative programming (Aos, Miller, &
Drake, 2006; Gaes, Flanagan, Motiuk, & Stewart, 1999; Lipsey
& Cullen, 2007; Mackenzie, 2006; Petersilia, 2004; Seiter &
Kadela, 2003). The evidence has been very consistent in
establishing that contact-driven supervision, surveillance, and
enforcement of supervision conditions have a limited ability to
6
Introduction
change offender behavior or to reduce the likelihood of
recidivism (Mackenzie, 2006). However, intensive supervision
programs with a clear treatment component do show a sizeable
impact on recidivism (Aos et al., 2006; Gaes et al., 1999;
Petersilia, 2004).
All of the strategies
MacKenzie identified as
effective focus on
dynamic criminogenic
factors, are skill-oriented,
are based on cognitivebehavioral models, and
treat multiple offender
deficits simultaneously.
MacKenzie (2006) recently summarized the “what works”
literature in corrections, with specific chapters on various
community corrections programs (e.g., life skills, cognitive
behavioral therapy, education, drug treatment, and intensive
supervision). She concluded that human service-oriented
programs were much more effective than those based on a
control or deterrent philosophy. In particular, there is growing
consensus that practices focusing on individual-level change,
including cognitive change, education, and drug treatment, are
likely to be more effective than other strategies, such as
programs that increase opportunities for work, reunite families,
and provide housing (see also Andrews & Bonta, 2006). All of
the strategies MacKenzie identified as effective focus on
dynamic criminogenic factors, are skill-oriented, are based on
cognitive/behavioral models, and treat multiple offender deficits
simultaneously. These conclusions are consistent with several
large meta-analyses of the evaluation literature (Andrews et
al., 1990; Aos et al., 2006; Lipsey & Cullen, 2007).
Selection of program type may be less important than proper
implementation of the program. Delivering a program in the
wrong context (i.e., intensive substance abuse treatment to
casual drug users) or poor implementation are common and
may explain most of the weak or null findings in the research
studies. Despite advances in knowledge and best practices,
studies of programs for offenders have documented persistent
problems in implementation and adherence to the fidelity of
evidence-based practice models (Lowenkamp, Latessa, &
Smith, 2006; Petersilia, 2004; Young, 2004). Additionally,
improperly implemented programs may be harmful. One recent
reentry program, Project Greenlight, was developed from
research and best practice models to create an evidence-based
reentry initiative which was evaluated with a random
assignment research design (Wilson & Davis, 2006). However,
the program participants performed significantly worse than a
comparison group on multiple measures of recidivism after one
year. The evaluators concluded that the New York program did
not replicate past best practice. Instead, Project Greenlight
7
Multi-site Evaluation of SVORI: Summary & Synthesis
modified past practice to fit institutional requirements, was
delivered ineffectively, did not match individual needs to
services, and failed to implement any post-release continuation
of services and support (Wilson & Davis, 2006; see also Rhine,
Mawhorr, & Parks, 2006; Visher, 2006; Marlowe, 2006). The
evaluators attributed the findings to a combination of
implementation difficulties, program design, and a mismatch
between participant needs and program content. A key
difficulty for Project Greenlight, as with many other communitybased reentry programs, was its lack of integration into an
overall “continuum of care” strategy that linked prison and
community-based treatment.
Effective rehabilitation
strategies have strong
program integrity,
identify criminogenic
factors, employ a multimodal treatment
approach, use an
actuarial risk
classification, and ensure
responsivity between an
offender’s learning style
and mode of program
delivery.
Another line of research has focused on identifying the
principles of effective treatment (as opposed to the substantive
content of the program) in assessing evidence-based practices
(e.g., Andrews & Bonta, 2006; Cullen & Gendreau, 2000).
MacKenzie (2006) summarizes this work, identifying five
principles of effective rehabilitation strategies. Specifically, she
notes that effective rehabilitation strategies have strong
program integrity, identify criminogenic factors, employ a
multimodal treatment approach, use an actuarial risk
classification, and ensure responsivity between an offender’s
learning style and mode of program delivery. One of the failings
of Project Greenlight was poor management of the program
according to these principles that help guide or maximize
program effectiveness (Andrews, 2006).
In her review of what works in reentry programming, Petersilia
(2004) discusses the striking disconnect between the published
‘what works’ literature and the efforts of governmental reentry
task forces to develop programs that are thought to improve
offender transitions from prison to the community. The goal of
most reentry programs is to develop a seamless transition from
prison to the community. However, the challenges in this
regard are enormous. Corrections departments and community
supervision agencies often have conflicting incentives, and
community-justice partnerships linking these organizations with
community groups face even larger hurdles. An important
barrier to effective reentry strategies in many communities is
the lack of information sharing between the criminal justice
system and the community because of institutional barriers and
privacy rules. Effective service delivery after release requires
coordinated actions by government agencies, non-government
8
Introduction
service providers, and the community to ensure that returning
prisoners do not fall through service gaps between agencies.
Yet, knowledge about how to develop and manage these
partnerships is lacking (Rossman & Roman, 2003).
The SVORI Multi-site
Evaluation was an
opportunity to test
whether coordinated
services provided in
response to assessment to
meet individual needs
could be implemented
and whether these
services would have
positive impacts on
criminal justice,
employment, health,
housing, and substance
use outcomes.
SVORI programs were developed and implemented by the
grantee agencies as these strands of research findings were
emerging. The programs were to provide a range of
coordinated services (based on needs and risk assessments)
that spanned incarceration and return to the community,
including services that focused on cognitive development.
Although the programs differed from site to site, as will be
further discussed, the overall focus of the SVORI initiative was
consistent with emerging recommendations at the time the
programs were developed and implemented. Thus, the SVORI
Multi-site Evaluation was an opportunity to test whether
coordinated services provided in response to assessment to
meet individual needs could be implemented and whether these
services would have positive impacts on criminal justice,
employment, health, housing, and substance use outcomes.
Juvenile Justice and Reentry
Juvenile offenders typically have serious and wide-ranging
deficits, including negative family influences and functioning,
mental health problems, low academic functioning, and high
rates of substance use. Juvenile offenders often have unmet
mental health needs, as illustrated by a rate of mental health
disturbance 2 to 3 times as high as the general adolescent
population (Grisso, 2004). It is estimated that 80% of juvenile
offenders suffer from minor mental health problems, including
conduct disorder, attention-deficit disorder, and mood and
anxiety disorders (Cocozza & Skowya, 2000; Mears, 2001).
Together with mental health problems, juvenile offenders
commonly experience physical health problems, as well as
learning disorders (National Council on Disability, 2003). In
addition, the National Research Council and Institute of
Medicine (2001) found that delinquency was associated with
poor school performance, truancy, and leaving school at an
early age.
9
Multi-site Evaluation of SVORI: Summary & Synthesis
Juvenile offenders have
serious and wide ranging
deficits including
negative family influences
and functioning, mental
health problems, low
academic functioning,
and high rates of
substance use.
Juvenile reentry and
transition services can
serve as an opportunity to
intervene and reverse a
downward trajectory for
many youth.
Substance use is also common among juvenile offenders. For
example, 9% of juvenile offenders younger than age 18
reported having used alcohol, 15% reported having used illicit
drugs, and 23% reported having used both alcohol and drugs at
the time they committed the crime that led to their
confinement (Kazdin, 2000). Other common characteristics of
juvenile offenders include criminally involved parents
(Farrington, 1989), poor parent-child relationships, and
inadequate parental supervision (Hawkins et al., 1998; Lipsey &
Derzon, 1998). These factors, if not adequately addressed, can
lead to failure in school, work, and personal relationships,
essentially precluding drug-free and crime-free lives after
release from confinement. Altschuler and Brash (2004)
summarized the challenges that confront juvenile offenders
upon release from confinement, noting, “When underlying
factors that predispose or propel them toward offending
behavior are not addressed during incarceration and afterward,
the likelihood is great that young offenders will reoffend upon
release” (p. 75).
Juvenile reentry and transition services can serve as an
opportunity to intervene and reverse a downward trajectory for
many youth (Freudenberg, Daniels, Crum, Perkins, & Richie,
2005). The transitional phase of community reentry, which has
been considered to be from one month pre-release to 6 months
post-release, is an important time for juvenile offenders to
establish lifestyles that do not support delinquency and criminal
activity (Altschuler & Brash, 2004). Juvenile offenders often
encounter problems similar to those that adult offenders
encounter when reentering their communities, such as
establishing supportive familial and peer relations after release.
For example, juveniles frequently return to the same
environments and family structures that contributed to their
delinquency; moreover, they often return to their communities
with serious unmet needs that complicate their opportunities
for successful reentry (Bouffard & Bergseth, 2008; Chung,
Schubert, & Mulvey, 2007). Although similar obstacles confront
adult and juvenile offenders, it is important to understand the
unique role that reentry plays in the lives of juvenile offenders
after their release from correctional institutions.
The juvenile justice system was originally established with the
goals of promoting the development of troubled youth and
training youth for successful adulthood, as well as, to a lesser
10
Introduction
extent, punishing youth for their offenses (Steinberg, Chung, &
Little, 2004). Feld (1998) suggests that, in response to youth
delinquency, in recent years the contemporary juvenile court
has increasingly emphasized punitive sanctions and public
safety over rehabilitation. Youth who complete their time with
the juvenile justice system too often reenter their communities
with just as many, if not more, problems than they had when
they first entered the system (Steinberg et al., 2004).
Because of the growing populations and crowding in juvenile
confinement facilities resulting from these “get tough” policies,
the ever-increasing costs of confinement, and high recidivism
rates, the 1980s marked a period in which policy makers and
practitioners began to reconsider the issue of juvenile reentry
(Altschulter & Armstrong, 1994). In 1987 the Office of Juvenile
Justice and Delinquency Prevention sought to assess, test, and
disseminate information about effective reentry programming
for serious, violent, and chronic juvenile offenders. The result of
this effort was the development of the Intensive Aftercare
Program (IAP), a theoretical and research-based model that
promoted intensive case management, the assessment and
identification of risk and needs factors, individualized case
planning, intensive supervision and monitoring, the use of
sanctions and rewards, and coordinated community-based
services. Moreover, IAP recognizes the importance of involving
all actors in the juvenile justice system, including providers
from child-serving agencies, to develop and implement a
seamless system for providing reentry services (Altschulter &
Armstrong, 1994).
In the past several years, the literature on reentry services for
confined youth has grown (Abrams, Shannon, & Sangalang,
2008; Bouffard & Bergseth, 2008; Freudenberg et al., 2005;
Mears & Travis, 2004; Steinberg et al., 2004). Attention has
been given to the domains and areas in which youth experience
particular challenges during reentry. According to Altschuler
and Brash (2004), these domains and areas include family and
living arrangements, peer groups, mental and physical health,
education, vocational training and employment, substance use,
and leisure activities.
11
Multi-site Evaluation of SVORI: Summary & Synthesis
Lipsey (2009) conducted
a meta-analysis of
juvenile reentry programs
and found that
interventions that
provided a therapeutic
element, served high-risk
offenders, and were
implemented with
expertise were considered
most effective.
Lipsey (2009) conducted a meta-analysis of juvenile reentry
programs and found that interventions that provided a
therapeutic element, served high-risk offenders, and were
implemented with expertise were considered most effective.
Similarly, MacKenzie (2006) has argued that multisystemic
therapy, which is a community-based treatment program for
serious juvenile offenders, is most effective for serious
offenders who are reuniting with their families, because
therapists and case managers are present to facilitate the
transition. In their examination of a juvenile reentry program
that offers the mentoring component of transitional
coordinators to released juveniles, Bouffard and Bergseth
(2008) concluded that juveniles who have participated in this
structured reentry program, in which services and group
planning are major elements, are more likely to successfully
reintegrate into the community. After a short-term follow-up,
such juveniles were found to have lower rates of recidivism
than juveniles who did not receive any reentry services.
Although findings from these studies are encouraging, research
on juvenile aftercare and reentry has been predominated by
null findings for program effects, small sample sizes, and
implementation challenges (Bouffard & Bergseth, 2008). As
some scholars have asserted, the skills acquired in juvenile
correctional facilities will not be sustained unless they are
reinforced in the community and are highly relevant to the reallife setting and situations these youth will confront once they
return to their communities (Abrams, 2006; Steinberg et al.,
2004).
THE SERIOUS AND VIOLENT OFFENDER
REENTRY INITIATIVE
The emerging consensus of the need for integrated, needsbased reentry programming for adult and juvenile offenders to
reduce recidivism and promote public safety provided the
context for the federal government’s Serious and Violent
Offender Reentry Initiative. The evaluation framework shows
the SVORI logic model and the evaluation components
(Exhibit 1). The SVORI program model identifies SVORI
funding, technical assistance, and requirements as inputs that,
in combination with local resources in the sites (throughputs),
yield a set of services and programming (outputs) expected to
improve the intermediate and recidivism outcomes for SVORI
12
Introduction
Exhibit 1. SVORI program logic model and evaluation framework
Community Context
Offender Context
Population Characteristics
Criminal History
Mental & Physical Health
Substance Abuse
Education/Training/Work Experience
Family Ties
Population Characteristics
Unemployment Rates
Service Availability
Residential Stability
Post-release Supervision Structure
Inputs:
The SVORI
Federal Funding &
Other Resources
Technical Assistance
Federal Grant
Requirements
Throughputs
Local Partnership
Formation &
Functioning
State & Local
Resources
Outputs:
Implementation
In-Prison
Coordination/Supervision
Education/Training
Family Services
Health Services
Transition Services
Community
Coordination/Supervision
Education/Training
Family Services
Health Services
Transition Services
Post-Supervision
Outcomes
Offender
Community Involvement
Employment
Family Contact/Stability
Health/Mental Health
Housing
Recidivism
Substance Use
Supervision Compliance
Systems
Rearrest Rates
Reincarceration Rates
Systems Change
Community Reintegration
Activities
Implementation Assessment
Evaluation
Components
Impact
Evaluation
Cost-Benefit Analysis
participants, as well as improve the state and local systems
that provide the services and programs. Community and
individual participant characteristics influence these
throughputs, outputs, and outcomes.
SVORI was an outcomeor goal-oriented initiative
that specified outcomes,
or goals, that were to be
achieved by programs
developed locally.
The SVORI program model shows that SVORI was an outcomeor goal-oriented initiative that specified outcomes, or goals,
that were to be achieved by programs developed locally.
Criteria specified by the federal partners for the local programs
were the following:3
ƒ
3
Programs were to improve criminal justice, employment,
education, health (including substance use and mental
health), and housing outcomes.
In some cases, grantees asked for and received exceptions to these
criteria. For example, some programs were primarily post-release
programs, and age restrictions were sometimes lifted (e.g., for
programs targeting sex offenders).
13
Multi-site Evaluation of SVORI: Summary & Synthesis
Each program was
locally designed . . . .
Consequently, one
challenge for the
evaluation was to
characterize SVORI.
Programs were to include collaborative partnerships
between correctional agencies, supervision agencies,
other state and local agencies, and community and
faith-based organizations.
ƒ
Program participants were to be serious or violent
offenders.
ƒ
Program participants were to be 35 years of age or
younger.
ƒ
Programs were to encompass three stages of reentry—in
prison, post-release on supervision, and postsupervision.
ƒ
Needs and risk assessments were to guide the provision
of services and programs to participants.
Operating within these broad guidelines, each program was
locally designed along a variety of dimensions, including the
types of services offered, the focus on pre-release and postrelease components, and the types of individuals to be served.
Programs varied in terms of what was being provided, when,
and to whom. Grantees also identified the locations where the
program would be provided both pre- and post- release. Thus,
a SVORI program could be narrowly focused on a single
institution pre-release, serving participants who were returning
to a single community post-release, or it could be implemented
throughout the correctional (or juvenile justice) system serving
participants who were to be released statewide. A combination
of multiple (but not all) institutions and multiple (but not all)
communities was the modal configuration. Finally, because
services were to be delivered to individuals on the basis of their
specific needs and risk factors, individuals participating in a
SVORI program could receive different types and amounts of
services, depending upon individual needs.4 Consequently, one
challenge for the evaluation was to characterize SVORI.
4
14
ƒ
Specific details on the planned characteristics of individual programs
are available in the National Portrait of SVORI (Lattimore et al.,
2004). Also see Lattimore et al. (2005), Winterfield et al. (2006),
and Lindquist and Winterfield (2005) for information on the delivery
of services and programs by the SVORI programs, together with
information on barriers to implementation.
Introduction
The SVORI evaluation
was intended to answer
the following research
questions:
● Did SVORI lead to
more coordinated
and integrated
services among
partner agencies?
● Did SVORI
participants receive
more individualized
and comprehensive
services than
comparable, nonSVORI offenders?
● Did reentry
participants have
better recidivism,
employment, health,
and personal
functioning outcomes
than comparable,
non-SVORI
offenders?
● Did the benefits
derived from SVORI
programming exceed
the costs?
The impact evaluation included pre-release interviews
(conducted approximately 30 days before release from prison)
and a series of follow-up interviews (conducted at 3, 9, and 15
months post-release). Nearly 2,400 prisoners returning to
society—some of whom received SVORI programming and
some of whom received “treatment as usual” in their respective
states—were included in the impact evaluation. Five programs
(adult programs in Iowa, Ohio, Pennsylvania, and South
Carolina; juvenile program in South Carolina) were selected for
the cost-benefit study, to assess the extent to which program
benefits exceeded costs.
Sixteen programs were selected for the impact evaluation,
comprising 12 adult programs and 4 juvenile programs located
in 14 states (adult only unless specified): Colorado (juveniles
only), Florida (juveniles only), Indiana, Iowa, Kansas (adults
and juveniles), Maine, Maryland, Missouri, Nevada, Ohio,
Oklahoma, Pennsylvania, South Carolina (adults and juveniles),
and Washington.5
MULTI-SITE EVALUATION IMPACT SITES
The impact sites represented a set of programs diverse in
approach and geographically distributed. Although the resulting
programs were not randomly selected, the adult programs were
in states that, at year’s end 2003, incarcerated about 20% of
all adult state prisoners and supervised about 23% of all adult
state parolees in the United States.6
The impact sites were representative of all sites along many
dimensions, although they were purposively selected. As
expected, the impact sites did vary from the non–impact sites
with regard to the criteria used in the selection process. In
particular, the impact sites planned, generally, to have larger
enrollments (Exhibit 2); larger enrollments were true for both
adult and juvenile impact sites. Although discrepancies between
5
6
Site selection and other methodological aspects of the study are
described in The Multi-site Evaluation of SVORI: Methodology and
Analytic Approach (Lattimore & Steffey, 2009).
Estimates are based on data from the U.S. Bureau of Justice
Statistics’ Adults on Parole in the United States (Glaze & Palla,
2005) and Prisoners under the Jurisdiction of State or Federal
Correctional Authorities (Harrison & Beck, 2005). The 12 states had
an estimated prison population of 259,971 midyear 2004 (19.8% of
all state prisoners) and 154,532 individuals on parole at year’s end
2004 (22.9% of all individuals under state parole supervision).
15
Multi-site Evaluation of SVORI: Summary & Synthesis
Exhibit 2. Program sizes among impact and non–impact sites (as reported by program
directors)
Program size
Adult planneda
Fewer than 100
101–150
More than 151
Adult in 2006b
Fewer than 100
101–150
More than 151
Adult compared with plannedc
Fewer
About the same
More
Juvenile plannedd
Fewer than 100 planned
101–150
More than 151 planned
Juvenile in 2006e
Fewer than 100 enrolled
101–150
More than 151 enrolled
Juvenile compared with plannedf
Fewer than originally projected
About the same as projected
More than originally projected
Impact Sites
%
N
Non–impact
Sites
%
N
All Sites
%
N
—
25.0
75.0
0
3
9
59.0
17.9
23.1
23
7
9
45.1
19.6
35.3
23
10
18
—
33.3
66.6
0
4
8
51.3
12.8
35.9
20
5
14
39.2
17.7
43.1
20
9
22
50.0
25.0
25.0
6
3
3
50.0
23.7
26.3
19
9
10
50.0
24.0
26.0
25
12
13
25.0
50.0
25.0
1
2
1
67.7
12.9
19.4
21
4
6
62.9
17.1
20.0
22
6
7
—
50.0
50.0
0
2
2
54.8
25.8
19.4
17
8
6
48.6
28.6
22.9
17
10
8
50.0
50.0
0.0
2
2
0
33.3
36.7
30.0
10
11
9
35.3
38.2
26.5
12
13
9
a
Fifty-one programs reporting; source: 2003 program work plan review.
b
Fifty-one programs reporting; source: 2006 program director survey.
c
Fifty programs reporting; source: 2006 program director survey.
d
Thirty-five programs reporting; source: 2003 program work plan review.
e
Thirty-five programs reporting; source: 2006 program director survey.
f
Thirty-four programs reporting; source: 2006 program director survey.
expected and actual enrollment were similar for adult impact
and adult non–impact sites, discrepancies between expected
and actual enrollments were not as similar for juvenile impact
and non–impact sites: one third of juvenile non–impact sites
experienced enrollments that exceeded expectations, while the
four juvenile impact sites experienced enrollments that either
failed to meet or met expectations. As of March 2006, the adult
16
Introduction
impact sites had enrolled an average of 326 program
participants, in comparison with an average enrollment of 290
participants by the non–impact sites. The juvenile impact sites
had enrolled an average of 153 SVORI program participants, in
comparison with an average of 204 participants by the juvenile
non–impact sites.
Program directors for the impact sites also were more likely
than program directors in non–impact sites to report being
closer to full implementation in both the 2005 and 2006
program director surveys (Exhibit 3). Again, this discrepancy
between impact and non–impact sites was expected, because
likelihood of full program implementation was one of the
selection criteria for inclusion in the impact evaluation.
Additional comparisons of the characteristics of the impact
sites, non–impact sites, and all sites that were derived from the
surveys of SVORI program directors are provided in Appendix C
of Lattimore and Steffey (Lattimore & Steffey, 2009). Overall,
these tables reveal relatively few differences in distributions for
adult or juvenile programs with regard to program director
turnover, basic program characteristics, targeted outcomes,
pre-release and post-release service provision, agency
involvement and contributions, stakeholder support and
resistance, and pre-release and post-release geographic
targeting. There were differences in expected pre-release and
post-release service enhancements, which may have been
associated with anticipated strength of implementation.
The methodological approach to the evaluation is described
briefly in the next chapter. Subsequent chapters summarize the
characteristics of the adult males, adult females, and juvenile
males who participated in the evaluation, and report findings on
service needs and service receipt both before and after release
from prison or detention. The penultimate chapter compares
outcomes between SVORI program participants and individuals
who did not participate in SVORI programs but may have
received services through other conduits. The final chapter
discusses the policy implications and need for future work.
17
Multi-site Evaluation of SVORI: Summary & Synthesis
Exhibit 3. Implementation status among adult impact and non–impact sites
Impact Sites
Program and status
%
N
a
Adult, fully operational in 2005
No
16.7
2
Yes
83.3
10
Adult, time to full implementation in 2005b
Less than 3 months
—
0
3–5 months
27.3
3
6–8 months
27.3
3
9–11 months
9.1
1
12 months or more
36.4
4
Adult, planned elements fully operational in 2006c
No
16.7
2
Yes
83.3
10
Juvenile, fully operational in 2005d
No
25.0
1
Yes
75.0
3
Juvenile, time to full implementation in 2005e
Less than 3 months
25.0
1
3–5 months
—
0
6–8 months
25.0
1
9–11 months
25.0
1
12 months or more
25.0
1
Juvenile, planned elements fully operational in 2006f
No
—
0
Yes
100.0
4
Non–impact Sites
%
N
37.5
62.5
15
25
32.7
67.3
17
35
20.0
3.3
23.3
13.3
40.0
6
1
7
4
12
14.6
9.8
24.4
12.2
39.0
6
4
10
5
16
7.9
92.1
3
35
10.0
90.0
5
45
15.2
84.8
5
28
16.2
83.8
6
31
27.6
10.3
27.6
13.8
20.7
8
3
8
4
6
27.3
9.1
27.3
15.1
21.2
9
3
9
5
7
3.3
96.7
1
29
2.9
97.1
1
33
a
Fifty-two programs reporting; source: 2005 program director survey.
b
Forty-one programs reporting; source: 2005 program director survey.
c
Fifty programs reporting; source: 2006 program director survey.
d
Thirty-seven programs reporting; source: 2005 program director survey.
e
Thirty-three programs reporting; 2005 program director survey.
f
Thirty-four programs reporting; source: 2006 program director survey.
18
All Sites
%
N
Methods
This chapter briefly describes the methods that were used for
the SVORI Multi-site Evaluation.7 Detailed information is
provided in Lattimore and Steffey (2009); for data collection
protocols and instruments, see the appendices to that
volume. Implementation evaluation data collection
procedures included document review with data extraction
and multiple surveys of the local SVORI program directors;
these procedures are described first. Next, the procedures
that were implemented to identify the impact sites and
potential impact evaluation participants are summarized. The
following section in this chapter describes the procedures for
collecting four rounds of interview data from respondents in
the impact sites. The final section describes the propensity
score approach that was used to address differences
between SVORI and non-SVORI groups on observed
characteristics.
IMPLEMENTATION DATA COLLECTION
The primary source of data for the implementation
assessment was four rounds of data collected from the
SVORI program directors. After extracting information
through reviews of the grant proposals and grantee work
plans, the evaluation staff conducted telephone interviews
with the program directors. This work began in August 2003
and concluded in October 2003. All 69 grantees responded,
and 88 separate programs were identified (one additional
program was subsequently identified for a total of 89).
Results, including descriptions of all programs, are
summarized in Lattimore et al. (2004).
7
Procedures for the collection of the cost-benefit data and the costbenefit analyses are described in Cowell et al. (2009).
19
Multi-site Evaluation of SVORI: Summary & Synthesis
The program directors were mailed hard-copy surveys in
March 2005. This survey collected additional information on
the planned structure of the SVORI program, enrollment to
date, and information on barriers and challenges to
implementation. Information on the types of programming
that would have been available for SVORI participants in the
absence of the SVORI program was also collected.
Responses were received from all 89 program directors,
although not every director responded to every question.
A second survey was mailed to the program directors in
March 2006. This survey collected updated information on
enrollment, as well as services provided, implementation,
and sustainability. Responses were received from 86 of the
89 program directors.
A final survey was e-mailed to the 89 program directors in
July 2007 to obtain information on ongoing reentry efforts in
their states after the conclusion of the SVORI grants. Data
were keyed by project staff, who also made follow-up
telephone and e-mail inquiries to increase response rates.
Responses were obtained from 52 of the 89 programs.
In addition to the program director surveys, which generated
descriptive data (self-reported) for all 89 SVORI programs,
two rounds of site visits were conducted with the subset of
programs included in the impact evaluation. The site visits
generated detailed information from a variety of key
stakeholders involved in SVORI (including line staff,
supervisors, and top administrators from the pre- and postrelease supervision agencies, service provider agencies, and
other key partners) and enabled the evaluation team to
more fully characterize program implementation, interagency
collaboration, and sustainability in the sites selected for the
impact evaluation.
IDENTIFICATION AND ENLISTMENT OF
IMPACT SITES
The strategy implemented to identify the impact programs
was to (1) review grant proposals and work plans to extract
program information, (2) conduct semistructured interviews
with all program directors to complete or clarify program
information and ability to participate in impact evaluation,
(3) complete site visits to a subset of sites, (4) analyze
20
Methods
collected information to generate a list of recommended
impact sites, and (5) obtain NIJ approval of proposed sites.
All data extracted from the documents and gathered from
interviews were stored in a database within the project
management information system.
The initial reviews identified a total of 89 distinct SVORI
programs and produced information on the SVORI program
target population, status of implementation, program
components and services, the capacity and willingness to
participate in the evaluation, the availability of sufficient
treatment and appropriate comparison populations, and
additional information on program goals and activities.8
Other information focused on agency involvement in SVORI,
management and oversight of the project, and plans to
conduct a local evaluation.
Once the initial data were gathered, the sites were examined
in accordance with the site selection criteria (Exhibit 4).
Implementation, target population size, comparison
respondent availability, and willingness to participate were
key factors that were considered.
Anticipated program
enrollment (greater than
100, unless the program
was in the same site as
another program with
enrollment greater than
100) and status of
program implementation
were the primary factors
used in developing the list
of potential impact sites
to be visited.
After the initial review of program type, enrollment, and
geographic targeting, staff narrowed down the 69 grantees
for site visits. Anticipated program enrollment (greater than
100, unless the program was in the same site as another
program with enrollment greater than 100) and status of
program implementation were the primary factors used in
developing the list of potential impact sites to be visited.
Twenty-nine grantees providing 39 separate programs in 21
states were selected for site visits. The primary purpose of
the site visits was to update information from the work
plans, gather information about the availability and quality of
administrative data, confirm program implementation
progress, assess site willingness to participate, and explore
opportunities for identifying comparison respondents (if the
site was not randomly assigning SVORI participation).
8
The 89 programs comprised 37 programs targeting juveniles only,
45 programs targeting adults only, and 7 programs targeting
both adults and juveniles (those younger than 18 years of age)
housed in adult facilities; for purposes of the evaluation, the
latter 7 programs were included with the adult programs.
21
Multi-site Evaluation of SVORI: Summary & Synthesis
Exhibit 4. Impact evaluation site selection criteria
Note: NIJ = National Institute of Justice.
The final list of sites proposed for the impact evaluation
incorporated a diversity of program types, geographic
regions, and corrections philosophies. This list was presented
to and discussed with NIJ in December 2003. Although the
original goal was to develop a set of sites at which to
interview about 4,000 respondents, case-flow analyses
conducted at the time of final site selection suggested that
the targets of 1,000 women and 1,000 juveniles would not
be attainable. Consequently, the goal was set to 3,000
respondents (1,500 SVORI participants and 1,500 nonSVORI comparison respondents). This total was to be
distributed into approximately 2,000 adult males, 500 adult
females, and 500 juvenile respondents (case flow was so
meager that enrollment of juvenile females was ended after
several months yielded only one SVORI program participant
and several comparison respondents).
Exhibit 5 identifies the sites and programs that were initially
identified for inclusion in the impact evaluation. One adult
site (Virginia) was dropped shortly after data collection
began because of logistical problems associated with the
identification and interviewing of respondents. The juvenile
22
Methods
Exhibit 5. Original programs selected for the impact evaluation
State
Grantee Agency
CO
Colorado Department of Corrections
a
FL
Florida Department of Juvenile Justice
IA
IN
KS
Iowa Department of Corrections
Indiana Department of Corrections
Kansas Department of Corrections
KS
Kansas Juvenile Justice Authority
ME
MEa
MD
MO
NV
OH
Maine Department of Corrections
Maine Department of Corrections
Maryland Department of Public Safety and
Correctional Services
Missouri Department of Corrections
Nevada Department of Corrections
Ohio Department of Rehabilitation and Corrections
OK
PA
Oklahoma Department of Corrections
Pennsylvania Department of Corrections
SC
SC
VAa
WA
South Carolina Department of Corrections
South Carolina Department of Juvenile Justice
Virginia Department of Corrections
Washington State Department of Corrections
Program
Colorado Affirms Reentry
Efforts (CARE)
Going Home
Iowa SVORI
Allen County SVORI
Shawnee County Reentry
Program (SCRP)
Kansas JJA Going Home
Initiative (GHI)
Maine Reentry Network
Maine Reentry Network
Re-Entry Partnership (REP)
Going Home-SVORI
Going Home Prepared
Community-Oriented Reentry
Program
PROTECT Oklahoma County
Erie, PA, Reentry Project
(EPRP)
SC Department of Corrections
SC DJJ Reentry Initiative
Going Home to Stay-VASAVOR
Going Home
Focus of Impact
Evaluation
Juveniles
Juveniles
(Dade County)
Adults
Adults
Adults
Juveniles
Adults
Juveniles
Adults
Adults
Adults
Adults
Adults
Adults
Adults
Juveniles
Adults
Adults
Subsequently dropped from the evaluation.
program in Maine was included in the original list of juvenile
impact sites but was dropped from the impact evaluation
because of insufficient case flow. In the end, 16 programs in
14 states were included in the impact evaluation. These sites
represented a set of programs that were diverse in approach
and geographically distributed. Although the resulting
programs were not randomly selected, the adult programs
were in states that, at year’s end 2003, incarcerated about
20% of all adult state prisoners and supervised about 23%
of all adult state parolees in the United States.9
9
Estimates are based on data from the U.S. Bureau of Justice
Statistics’ Adults on Parole in the United States (Glaze & Palla,
2005) and Prisoners Under the Jurisdiction of State or Federal
Correctional Authorities (Harrison & Beck, 2005). The 12 states
had an estimated prison population of 259,971 in midyear 2004
(19.8% of all state prisoners) and 154,532 individuals on parole
at year’s end 2004 (22.9% of all individuals under state parole
supervision).
23
Multi-site Evaluation of SVORI: Summary & Synthesis
The impact sites were representative of all sites along many
dimensions, although they were purposively selected. As
expected, the impact sites did vary from the non–impact
sites with regard to the selection criteria. In particular, the
impact sites generally planned to have larger enrollments,
and program directors for the impact sites were more likely
than program directors for non-impact sites to report being
closer to full implementation in both the 2005 and 2006
program director surveys.
Once sites were selected, the process of recruiting and
complying with requirements for conducting research in
these agencies’ facilities began. Recruitment included letters
from NIJ administrators and the evaluation principal
investigators, telephone contacts, and site visits. Memoranda
of understanding or research agreements were signed with
all agencies. Once negotiated research agreements were in
place, the next step was to develop evaluation plans for each
site, with a particular goal of identifying appropriate
comparison respondent populations. This process is
described in the next section.
IDENTIFICATION OF COMPARISON
RESPONDENT POPULATIONS
Two pathways were identified for inclusion in SVORI
programs: (1) random assignment to SVORI programming or
standard programming after a decision to participate in
SVORI was made by the offender and (2) program and
offender determination of SVORI program eligibility and
participation. For the first pathway (two adult sites), those
randomly assigned to standard programming constituted the
pool of potential comparison respondents. For the second,
evaluation team members worked with local personnel to
identify the site-specific SVORI eligibility criteria and to
establish procedures for selecting a comparison group. In
most cases, the comparison respondents were offenders who
would have been eligible for (i.e., offered) SVORI if they had
been in a facility that offered the SVORI program or if they
had planned to return to a community with a post-release
SVORI program.
24
Methods
The SVORI Multi-site
Evaluation took an
“intent- to-treat”
approach.
The SVORI Multi-site Evaluation took an “intent-to-treat”
approach with respect to the classification of respondents as
SVORI participants or non-SVORI comparison respondents.
Practically, this meant that an individual was classified as
SVORI or non-SVORI, depending upon whether he or she
was enrolled in a SVORI program at any time during the
period between the receipt of the case from the site and the
date the case was fielded.
IMPACT DATA COLLECTION: INTERVIEWS
AND ADMINISTRATIVE DATA
The data collection consisted of four in-person interviews
with offenders (approximately one month before release and
3, 9, and 15 months after release), supplemented by
administrative data from state departments of correction,
probation and parole, and juvenile justice, as well as from
the National Center for Crime Information (NCIC). Drug tests
(oral swabs) were conducted at the 3- and 15-month
interviews with participants who were in the community (i.e.,
not in a correctional or treatment facility) at the time of the
interview and provided a separate consent for the tests.
Offender Interviews
Eligible respondents (both SVORI and non-SVORI) were
identified on a monthly (or more frequent) basis during a 16month Wave 1 (pre-release) interviewing period (July 31,
2004, through November 30, 2005). All interviews were
conducted in private settings, using computer-assisted
personal interviewing (CAPI), by experienced RTI field
interviewers who had completed extensive training.
The Wave 1 interviews were conducted in prisons, juvenile
detention facilities, and halfway houses about 30 days before
the respondent’s expected release. This interview lasted
approximately 1.5 hours and was designed to obtain data on
respondents’ experiences and receipt of services during
incarceration, as well as to document respondents’
immediate post-release plans. No compensation was
provided for the Wave 1 interviews.
Waves 2 through 4 follow-up interviews were conducted at 3,
9, and 15 months after release. The follow-up interviews
lasted approximately 1.5 hours and covered topics such as
25
Multi-site Evaluation of SVORI: Summary & Synthesis
housing, employment, education, family, peer relationships,
community involvement, physical and mental health,
substance use, crime and delinquency, supervision, service
needs, and service receipt. The follow-up interviews were
conducted in the community or, for those reincarcerated, in
prison or jail. For interviews conducted in the community,
respondents were compensated $35 for the 3-month
interview, $40 for the 9-month interview, and $50 for the
15-month interview. At the final interview, respondents were
paid an additional $50 if they completed all four interviews.
In addition, respondents were paid an extra $5 at each
follow-up wave if they called a toll-free number to schedule
their interviews. As the 15-month interview was fielded, the
original protocol with respect to compensation was adjusted
so that, wherever agreements could be negotiated with
corrections departments and local jails, incarcerated
participants were compensated.
A total of 4,354 cases were fielded at Wave 1, resulting in
2,583 completed Wave 1 interviews. Of these, 192
respondents were not released during the time Wave 2 (3
month post-release) interviews were being conducted, and
these respondents were retrospectively declared ineligible for
the evaluation. As a result, the final tally for the three
demographic groups was 2,391 completed Wave 1 interviews
with evaluation-eligible respondents. The remaining cases
included 635 cases that were determined to be ineligible for
the evaluation, 718 cases that were released before an
interview could be scheduled, 370 refusals, and 48 other
noninterviews (due primarily to the inaccessibility of the
respondents).
The numbers of completed interviews (all waves) by SVORI
status and demographic group are shown in Exhibit 6. The
distributions of these cases by site are provided in Appendix
Exhibit A-1.
26
Methods
Exhibit 6. Completed interviews by wave, SVORI status, and demographic group
State
Adult males
Adult females
Juvenile males
Total
Wave 1
NonSVORI SVORI
863
834
153
204
152
185
1168
1223
Wave 2
NonSVORI SVORI
529
455
110
134
105
131
744
720
Wave 3
NonSVORI SVORI
565
470
119
134
108
131
792
735
Wave 4
NonSVORI SVORI
582
531
124
152
107
141
813
824
Note: Wave 1 = 30 days pre-release; Wave 2 = 3 months post-release; Wave 3 = 9 month post-release; Wave 4 =
15 months post-release.
Oral swab drug tests were conducted in conjunction with 3and 15-month interviews held in the community (i.e., not in
prisons, jails, or treatment facilities). Respondents were
provided an additional $15 if they consented to provide an
oral swab. The chosen test was a six-panel oral fluid screen
for amphetamines, cannabinoids, cocaine,
methamphetamines, opiates, and phencyclidine. All positive
findings were confirmed by gas chromatography/mass
spectrometry.
Impact: Administrative Data
The evaluation requested official criminal records data to
supplement the self-reported interview data, particularly with
respect to measures of criminal history and recidivism. The
sources of data were (1) state departments of corrections,
departments of juvenile justice, and probation and parole
agencies and (2) the NCIC. State agencies provided data on
return to prison after prisoner release, as well as information
on performance during post-release parole or probation. The
NCIC provided data on arrests, as well as information on
convictions and reincarcerations for some states.
PROPENSITY SCORE MODELS
Propensity score models were used to address potential
selection bias due to the quasi-experimental design (see
Rubin, 2006, for a collection of seminal papers in propensity
score modeling; see D’Agostino, 1998, for an accessible
27
Multi-site Evaluation of SVORI: Summary & Synthesis
tutorial).10 Propensity score models use observed
characteristics to model the likelihood that an individual with
those characteristics will be selected (or assigned) to the
intervention. The purpose is to identify a set of parameters
that are then used to estimate the probability of assignment
to the intervention for each individual in a study. These
probabilities (p-hats or p̂ ) are then used either (1) to
establish probability strata (or bins) within which
respondents are grouped together by similar probabilities of
receiving the intervention, (2) as weights in the outcome
models, or (3) as matching variables by which respondents
in the intervention group are matched to respondents in the
comparison group who have similar p̂ . The success of the
propensity score model estimation is judged by the
effectiveness of the strata or weights to reduce differences
between the treatment and control groups on observed
characteristics or, in the common terminology, “to achieve
balance” between the two groups.
Initial outcome models were examined with the use of the
stratification, or binning, approach, but the final outcome
models were estimated with the use of the weighting
approach because it greatly simplified the presentation of
10
28
Propensity scoring methods are not without limitations. For
example, use of propensity scores can only adjust for included
covariates (Glynn, Schneeweiss, & Sturmer, 2006; Rosenbaum &
Rubin, 1983). Unlike randomization, which tends to balance
treatment and control groups on observed and unobserved
covariates, use of propensity scores balances only on observed
confounding covariates. The failure to include unobserved
covariates can lead to biased estimates of treatment effects.
However, if many of the covariates believed to be related to
treatment assignment are measured, propensity score
approaches (i.e., matching, stratification, regression adjustment)
should yield consistent and approximately unbiased estimates of
treatment effects (D'Agostino, 1998; Rosenbaum & Rubin,
1983). A second limitation is that propensity score approaches
work better in larger samples; in studies with small samples,
substantial imbalances of covariates may be unavoidable (Rubin,
1997). However, this is also true of randomized experiments and
is not limited to propensity score methods. A third possible
limitation is that included covariates that are strongly related to
treatment assignment and only weakly correlated with the
outcome are treated the same as covariates that are strongly
related to both treatment assignment and outcome (Rubin,
1997). This might be considered a limitation because including
irrelevant covariates can reduce efficiency. Rubin (1997) notes,
however, that the potential biasing effects of failing to control for
weakly correlated covariates are worse than the potential loss of
efficiency from including them.
Methods
findings.11 Using the weighting approach allowed the
estimation of one set of outcome models for each
demographic group. Presenting findings by strata would have
multiplied the number of models and results to be presented
by the number of strata. For example, if adult male
respondents were assigned to one of five strata, differences
in outcomes would have to be assessed within each strata
increasing the number of models by a factor of five.
The success of the propensity score model estimation is
judged by the effectiveness of the strata or weights in
reducing differences between the treatment and control
groups on observed characteristics or, in the common
terminology, to achieve balance between the two groups.
Two ways of checking for balance are (1) to examine tstatistics comparing group means or (2) to examine
standardized differences between the two groups (see, e.g.,
Rosenbaum & Rubin, 1985). Both approaches were used and
indicated that the propensity score weights generated good
balance for the data.
The propensity score approach is useful only if it produces
adequate overlap in the p̂ between groups. The goal is to
develop scores that, for example, can be used to sort
individuals into strata, where the probability of assignment
to the intervention is similar. Once individuals are assigned
to strata on the basis of their p̂ , the strata should contain
individuals from both groups—otherwise there is no
comparison between groups. The propensity score models
produced p̂ distributions with considerable overlap between
SVORI and non-SVORI respondents for all three demographic
groups.
Item missingness was relatively rare in the data, but
imputation procedures were employed so that no
observations would have to be excluded from the outcome
analyses because of item missingness. Logit models to
generate the probability of assignment to SVORI [p(SVORI),
or p(S)] were estimated within the framework of SAS 9.1.3
11
Preliminary results showed that population average treatment
effects estimated by combining results from the analyses based
on strata for the adult male groups were nearly identical to those
derived from the weighted models—as would be expected.
Results were also similar for the adult female and juvenile males
groups; those results are not presented here.
29
Multi-site Evaluation of SVORI: Summary & Synthesis
PROC MI and PROC MIANALYZE for each of the three
demographic groups (adult males, adult females, and
juvenile males). These SAS procedures accommodated item
missingness by imputing values for missing data. A two-step
imputation procedure was used within PROC MI, in which
(1) a Monte Carlo procedure (MCMC) was employed to
impute values until the data set reached a pattern of
monotone missingness and then (2) regression was
employed to impute the remaining values (see SAS Institute
Inc., 2004; Allison, 2001). The independent variables for the
propensity score models included only variables that
reflected the values of measures before program assignment
(effectively, pre-incarceration).
Population average treatment effect weights were generated
from the propensity scores and applied to the observations
for analyses. SAS 9.1.3 Proc Survey Means was used to
generate weighted means; Proc Survey Logistic and Proc
Survey Regression were used to generate tests of
significance (see Lattimore & Steffey, 2009).
30
Respondent
Characteristics
This chapter uses data from the Wave 1 interviews to profile
the respondents that were included in the Multi-site Evaluation
and compares the SVORI respondents with the non-SVORI
respondents.
Overall, on average, the participants in the study reported
extensive criminal histories, lifetime substance use, histories of
substance use treatment, low educational attainment, and weak
employment histories. The adult male respondents are
described first, followed by the adult female respondents and
juvenile male respondents. The final section in this chapter
assesses similarities and differences among the three groups.
ADULT MALE SVORI EVALUATION
PARTICIPANTS
The following description of the adult male study participants is
adapted from Prisoner Reentry Experiences of Adult Males:
Characteristics, Service Receipt, and Outcomes of Participants
in the SVORI Multi-site Evaluation (Lattimore et al., 2009).12
The average adult male study respondent was 29 years of age;
about half were black and one third were white. Before
incarceration, most respondents reported that they had lived in
a house or apartment that belonged to someone else, and
about two thirds said that they were involved in a steady
relationship during that time. About 60% reported having
12
Preliminary findings were presented in the Pre-release
Characteristics and Service Receipt Among Adult Male Participants
in the SVORI Multi-site Evaluation (Lattimore, Visher, & Steffey,
2008).
31
Multi-site Evaluation of SVORI: Summary & Synthesis
children younger than 18 years old, and those who had children
reported having an average of more than two. Nearly all
reported having family members and friends who had been
convicted of a crime or had problems with drugs or alcohol.
About 60% reported that they had a high school diploma or
GED credential.
Substance Use and Physical and Mental Health
ƒ
Nearly all respondents reported lifetime use of alcohol
(96%) and marijuana (93%), whereas more than half
(55%) reported cocaine use.
ƒ
Of those who had ever used drugs, about two thirds
reported having used one or more illicit drugs during the
30 days before their incarceration, with about 52%
reporting marijuana use and 24% reporting cocaine use.
ƒ
Most respondents reported few physical health
problems, and most described their mental health status
at the time of the pre-release interview as excellent or
very good.
Employment History and Financial Support
ƒ
Most study participants (90%) reported having worked
at some point during their lifetimes, and about two
thirds reported working during the 6 months before
prison.
ƒ
Of those working during the 6 months before prison,
about three quarters described their most recent job as
a permanent job for which they received formal pay.
ƒ
Nearly half of the respondents reported supplementing
their legal income with income from illegal activities,
with those who had no job before prison more likely to
report illegal income.
Criminal History
32
ƒ
The respondents reported an average age at first arrest
of 16 and an average of 12 arrests.
ƒ
Most respondents had been previously incarcerated, and
about half had been detained in a juvenile facility.
ƒ
At the time of the interview, respondents reported an
average length of incarceration of more than 2 years.
ƒ
Most respondents indicated that family members had
served as an important source of support during their
incarceration.
Respondent Characteristics
Differences Between SVORI and Non-SVORI
Although the SVORI and comparison respondents were similar
on many of several hundred measures examined, they differed
on a few measures:
ƒ
SVORI respondents were more likely to be black (57%,
as opposed to 50%) and less likely to be white than
non-SVORI comparison respondents (32%, as opposed
to 37%).
ƒ
Non-SVORI respondents were significantly more likely
than SVORI respondents to indicate symptoms of
hostility and psychosis on the mental health subscales.
ƒ
Self-reports on “ever using” drugs indicated somewhat
higher usage among the non-SVORI respondents (96%
of non-SVORI, as opposed to 94% SVORI reported ever
using at least one illegal drug). Although this overall
difference was not statistically significant (at alpha =
.05), non-SVORI respondents were significantly more
likely than SVORI program participants to report ever
using most illegal drugs. The two groups did not differ,
however, with respect to reported use in the 30 days
before the instant incarceration.
ƒ
SVORI respondents were somewhat less likely to have
been employed before incarceration, although the
difference was small (89%, as opposed to 92%).
ƒ
SVORI respondents were less likely to be in prison for a
parole violation (27%, as opposed to 35%).
ƒ
SVORI respondents were more likely to be serving time
for a drug crime (36%, as opposed to 31%).
ƒ
SVORI respondents had spent more time in prison
during the current incarceration (2.8 years, as opposed
to 2.3 years).
ƒ
SVORI respondents reported more disciplinary
infractions (47%, as opposed to 41% reported two or
more) and were more likely to report having been
placed in administrative segregation (45%, as opposed
to 40%) than the non-SVORI respondents, which may
be associated with their longer lengths of stay.
ADULT FEMALE SVORI EVALUATION
PARTICIPANTS
The following description of the adult female study participants
is adapted from Prisoner reentry experiences of adult females:
33
Multi-site Evaluation of SVORI: Summary & Synthesis
Characteristics, service receipt, and outcomes of participants in
the SVORI Multi-site Evaluation (Lindquist et al., 2009).
The adult female respondents were an average of 31 years old
at the time of the pre-release interview, with nearly equal
numbers self-identifying as white (44%) and black (41%).
While 41% of the women reported living in their own homes
during the 6 months before incarceration, more than one fifth
were homeless, living in a shelter, or without a set place to live.
The vast majority of women were mothers; more than half of
those with minor children reported that they had primary care
responsibilities before incarceration. Approximately 62% of the
women reported having a high school diploma or GED.
Substance Use and Physical and Mental Health
ƒ
Nearly all women reported having used alcohol (96%)
and marijuana (90%) during their lifetimes, and 75%
reported cocaine use. More than two thirds (68%) of the
women reported having used one or more illicit drugs
during the 30 days before incarceration.
ƒ
Women reported many physical and mental health
problems; at the time of the pre-release interview,
fewer than half rated their physical health and less than
one third rated their mental health as excellent or very
good.
ƒ
Half of the women reported receiving treatment for
mental health problems before their instant
incarceration.
Employment History and Financial Support
ƒ
Most women (95%) reported having worked at some
point during their lifetimes; more than half reported
working during the 6 months before prison.
ƒ
Of those who worked during the 6 months before prison,
about three quarters reported that their most recent job
was permanent and that they received formal pay.
ƒ
Nearly half of the women reported receiving income
from illegal activities, with those lacking a job before
prison being more likely to report illegal income.
Criminal History
ƒ
34
The women reported an average of 11 arrests, with the
first arrest occurring, on average, at 19 years of age.
Respondent Characteristics
ƒ
Nearly all women reported at least one previous
incarceration; one third had been detained in a juvenile
facility.
ƒ
At the time of the interview, women reported an
average length of incarceration of less than 2 years.
Differences Between SVORI and Non-SVORI
The results from the pre-release interviews show that the
SVORI and non-SVORI groups were similar on most background
characteristics. There were some differences in some important
measures, including the following:
ƒ
SVORI respondents were more likely to have a 12th
grade education or GED than comparison respondents
(77%, as opposed to 55%).
ƒ
Self-reports on lifetime drug use indicated higher usage
of heroin (27%, as opposed to 18%) and amphetamines
(41%, as opposed to 31%) among the SVORI
respondents while self-reports on use 30 days before
incarceration indicated higher cocaine use among the
non-SVORI respondents (64%, as opposed to 50%).
ƒ
SVORI respondents were more likely to report having no
physical health limitations than the non-SVORI
respondents.
ƒ
Several measures indicated that SVORI women had
better mental health than the non-SVORI women.
ƒ
Among respondents who worked during the 6 months
before incarceration, SVORI respondents were more
likely to describe their most recent job as a permanent
job (82%, as opposed to 69%) and one for which they
received formal pay (91%, as opposed to 68%).
ƒ
SVORI respondents had spent more time in prison
during the current incarceration (2.2 years, as opposed
to 1.3 years).
ƒ
SVORI respondents reported more disciplinary
infractions (42%, as opposed to 29% reported two or
more) than the non-SVORI respondents, which may be
associated with their longer lengths of stay.
JUVENILE MALE SVORI EVALUATION
PARTICIPANTS
The following description of the juvenile male study participants
is adapted from Reentry experiences of confined juvenile
offenders: Characteristics, service receipt, and outcomes of
35
Multi-site Evaluation of SVORI: Summary & Synthesis
juvenile male participants in the SVORI Multi-site Evaluation
(Hawkins et al., 2009).
The average age of the juvenile male respondents was 17; 54%
were black, 20% were white, and 20% were Hispanic. At their
pre-release interview, most respondents reported that they
were currently in school. In the school year before their
confinement, less than half reported that they were regularly
attending school, and nearly all respondents reported that they
had been suspended or expelled from school during their
lifetimes.
Most respondents reported that, before confinement, they were
living in a house or apartment that belonged to someone else
(including parents’ house or apartment). Respondents most
frequently reported their natural mothers as the primary person
who raised them and the person with whom they had lived the
longest. Nearly all agreed or strongly agreed that they felt close
to their families and wanted their families to be involved in
their lives. More than three quarters of respondents reported
that they had family members who had been convicted of a
crime or had been incarcerated. More than half of respondents
reported that they had family members who had had problems
with alcohol or drugs. A large majority of respondents reported
that they had friends who had been convicted of a crime, had
been incarcerated, or who had had problems with alcohol or
drugs.
Substance Use and Physical and Mental Health
36
ƒ
Nearly all respondents (91% of SVORI and 83% of nonSVORI respondents) reported that they had used
alcohol. The average age at first use was 12.
ƒ
A large majority of respondents (>85%) reported that
they had used marijuana. The average age at first use
was 12.
ƒ
Most respondents reported that they had used alcohol or
other drugs in the 30 days before confinement.
ƒ
About half of all respondents reported that they had
received treatment for a substance use or mental health
problem at some point during their lifetimes.
ƒ
Most respondents rated their current physical health as
excellent or very good. More than half of all respondents
described their mental health status as excellent or very
good.
Respondent Characteristics
Employment History and Financial Support
ƒ
Nearly half of all respondents (43% of SVORI and 51%
of non-SVORI) reported having worked at some time
before confinement. More than one third reported that
they were employed in the 6 months before
confinement.
ƒ
Of those working in the 6 months before confinement,
about half described their most recent job as
permanent.
ƒ
The majority of respondents reported that they received
financial support from their family. About one third of
respondents reported that they supported themselves by
illegal income.
Delinquency History and Current Offense
ƒ
On average, respondents were 13 years old at the time
of first arrest, had been arrested about six times, and
had been adjudicated about three times.
ƒ
Nearly all respondents (>88%) previously had been
ordered to a juvenile correctional facility.
ƒ
In the 6 months before confinement, about three
quarters of respondents reported that they had engaged
in violent behavior, and nearly two thirds reported that
they had been victims of violence.
ƒ
More than 10% of respondents reported having been a
member of a gang.
ƒ
Nearly half of respondents reported that they were
currently confined for a violent crime.
ƒ
At the time of their pre-release interview, respondents
reported an average length of confinement of more than
one year.
Differences Between SVORI and Non-SVORI
Although the SVORI and non-SVORI comparison respondents
were similar on many of the several hundred measures, they
differed significantly on a few measures:
ƒ
SVORI respondents were older (17, as opposed to 16.7
years) and less likely to be white (14%, as opposed to
24%) than comparison respondents.
ƒ
SVORI respondents were more likely than comparison
respondents to report that they had family members
who had been convicted of crimes (87%, as opposed to
76%).
37
Multi-site Evaluation of SVORI: Summary & Synthesis
ƒ
SVORI respondents were more likely than comparison
respondents to report that they had received formal pay
at their most recent job (52%, as opposed to 31%).
ƒ
Comparison respondents reported better physical health
than SVORI respondents (55.0, as opposed to 53.4 on
the 12-Item Short Form Health Survey [SF-12] physical
health scale).
ƒ
SVORI respondents were more likely than comparison
respondents to indicate symptoms of phobic anxiety and
psychoticism.
ƒ
SVORI respondents were more likely than comparison
respondents to report that they had at some time used
alcohol (91%, as opposed to 83%) and hallucinogens
(30%, as opposed to 19%).
ƒ
SVORI respondents were less likely than comparison
respondents to be currently confined for a drug (11%,
as opposed to 19%) or public-order crime (20%, as
opposed to 32%).
ƒ
On average, SVORI respondents had fewer prior terms
of confinement to a juvenile correctional facility than
comparison respondents (2.97, as opposed 3.65) but
were significantly more likely to report that they had at
some time been detained for more than 24 hours at one
time (60%, as opposed to 48%).
COMPARISON OF ADULT MALE, ADULT
FEMALE, AND JUVENILE MALE SVORI
EVALUATION RESPONDENTS
Perhaps the most interesting difference between the juvenile
males and the male and female adults is that the juveniles were
substantially younger than the adults at the self-reported age
of initiation into substance use and criminal behavior. Exhibit 7
shows that the juvenile males were arrested for the first time at
an average age of 13 years in comparison to 16 and 19 years
for the adult males and females. The adult females reported
being about 3 years older at first arrest and at incarceration
than the men.
38
Respondent Characteristics
Exhibit 7. Average ages at first arrest, instant incarceration, and release
16
First arrest
19
13
27
Incarceration
30
16
29
Release
32
17
0
5
10
15
20
25
30
35
Years
Adult males
Adult females
Juvenile males
Exhibit 8 shows the average age of onset for alcohol,
marijuana, and cocaine (the most commonly reported
substances) use. Again, the juvenile males reported
substantially lower ages at first use for each of these drugs—
reporting initial use of alcohol and marijuana at age 12 in
comparison to age 14 or 15 for the adult respondents.
Similarly, the initial age of use of cocaine was also substantially
lower—15, as opposed to about 20 for the adults. However, the
juvenile males were much less likely than the adults to report
cocaine use—as can be seen in Exhibit 9. The adult females
were most likely to report cocaine use (74% SVORI; 75% nonSVORI) compared with about half of the adult males (53%
SVORI; 58% non-SVORI) and about one quarter of the juvenile
males (25% SVORI; 26% non-SVORI).
39
Multi-site Evaluation of SVORI: Summary & Synthesis
Exhibit 8. Age at first use of alcohol, marijuana, and cocaine
19
First cocaine
20
15
14
First marijuana
15
12
14
14
First alcohol
12
0
5
10
15
20
25
Years
Adult males
Adult females
Juvenile males
Note: Differences between adult male and female ages and juvenile male ages were significant at p < 0.05.
Exhibit 9. Percentages of respondents reporting use of alcohol, marijuana, and cocaine, by
SVORI program status and demographic group
100%
96%
97% 94%
92%
97%
80%
60%
95%
91%
88%
83% 85%
75%
74%
53%
91%
89%
58%
40%
25%
26%
Juvenile male,
SVORI
Juvenile male,
non-SVORI
20%
0%
Adult male,
SVORI
Adult male,
non-SVORI
Adult female,
SVORI
Alcohol
40
Adult female,
non-SVORI
Marijuana
Cocaine
Respondent Characteristics
All three demographic groups reported numerous prior arrests
and convictions. Exhibit 10 shows the average number of prior
arrests, convictions and incarcerations or detentions for each
demographic group. The juvenile males have substantial
histories compared with their adult counterparts: The men
were, on average, 16 at their first arrest and 27 at the time of
the instant incarceration, a span of 11 years in which to
accumulate an average of 14 prior arrests and 6 prior
convictions (ignoring time off the street)—somewhat more than
one arrest and about a half a conviction a year.13
Exhibit 10. Numbers of prior arrests and convictions, by demographic group
Number
16
14
Adult males
14
Adult females
12
Juvenile males
12
10
7
8
6
5
6
3
4
2
0
Number of arrests
Number of convictions
The juvenile males, on average, took about 3 years (16 years
at instant incarceration minus 13 years at first arrest) to
accumulate an average of 7 arrests and 3 convictions—about 2
arrests and one conviction per year or twice the rate of the
adult men. The adult women reported accumulating arrests and
convictions at roughly the same rate as the adult men—12
arrests and 5 convictions over an 11-year period (30 years at
instant incarceration minus 19 years at first arrest).
13
The average numbers of arrests reported here are the averages
calculated using all values as reported by the respondents; in some
cases, the evaluation has reported prior arrests that reflected the
recoding of extreme values at the 95th or 99th percentile. The
average numbers of prior arrests when values were capped at the
95th percentile were 13 for the adult males, 11 for the adult
females, and 6 for the juvenile males.
41
Multi-site Evaluation of SVORI: Summary & Synthesis
The percentages reporting that they were currently serving
time for person/violent, drug, and property offenses are shown
in Exhibit 11. As can be seen, the juvenile males were most
likely to report having a person/violent offense, as well as a
property offense. Women were more likely than the men to
report having a property offense.
Exhibit 11. Percentages reporting person/violent, drug, and property offenses associated
with instant incarceration, by demographic group
Adult males
60%
Adult females
46%
50%
48%
Juvenile males
41%
41%
40%
30%
33%
29%
29%
25%
20%
15%
10%
0%
Person/violent
Drug
Property
There were also differences among the groups in physical and
mental health—with the women scoring worse on functioning
and psychopathology scales than the men or juvenile males.
Exhibit 12 shows the SF-12 physical and mental health scores
for the three demographic groups. Higher scores indicate better
functioning; the scales are normed at 50 (Ware, Kosinski, &
Keller, 1996). As can be seen, the women had worse scores on
physical and mental health scales than the men and the
juvenile males. Exhibit 12 also shows the average Symptom
Assessment–45 Questionnaire (SA-45) Global Severity Index
scores (with higher scores indicating greater psychopathology;
Strategic Advantages, 2000). Again, the women have worse
scores than the men and juvenile males.
In the next section, the self-reported service needs and service
receipt are discussed.
42
Respondent Characteristics
Exhibit 12. 12-Item Short-Form mental and physical health scales, by demographic group
Adult males
100
Adult females
79
Juvenile males
80
67
60
49
50
44
54
49
63
54
40
20
0
SF-12 Mental Health
SF-12 Physical Health
Global Severity Index
Note: SF-12 = 12-Item Short Form Health Survey mental and physical health scales.
43
Service Needs and
Service Receipt
This chapter summarizes what the respondents reported about
their service needs and service receipt at each wave of
interviews and compares the self-reports of SVORI and nonSVORI respondents. Responses are reported for important
service items and domain-specific “bundle scores” that identify
the proportion of services a respondent reported needing or
receiving. Appendix Exhibits A-2 through A-7 contain the details
on the expressed service needs and reported service receipt,
providing the weighted means for individual items and bundle
scores by demographic group and data collection wave.
SVORI service bundle
scores were developed to
summarize service needs
and service receipt by
summing indicators of
needs and receipt in
multiple domain areas.
Bundle scores were developed from the interview data to
summarize respondents’ needs in the domains of transition,
health, employment/education/skills, domestic violence, and
child services (which was calculated only for respondents with
children). Scores for each individual were generated by
summing one/zero indicators for whether the individual did or
did not report needing each of the items within a bundle; this
sum was divided by the number of items in the bundle and
multiplied by 100.14 At the individual-respondent level, this
bundle score can be interpreted as the percentage of the
services in the bundle that the individual reported needing
(Winterfield et al., 2006). At the group level, the bundle score
indicates the average percentage of services needed. Similar
scores were developed to summarize service receipt. An
additional bundle—coordination services—was added to the set
14
Responses were “a lot,” “a little,” and “not at all.” The results
presented here are based on responses recoded to “some” or “not
at all.”
45
Multi-site Evaluation of SVORI: Summary & Synthesis
of service receipt bundles. Exhibit 13 shows the items that were
included in the service needs and receipt bundles.
Exhibit 13. Items included in service need and service receipt bundles
Need
Receipt
Coordination Service Bundle
a needs assessment
meeting with a case manager
collaboration with someone to reintegrate
currently on probation or parolea
needs assessment specific for releaseb
reentry plan developedb
help accessing child welfare case workerc
meeting with child welfare case workerc
Transition Service Bundle
legal assistance
legal assistance
financial assistance
financial assistance
public financial assistance
public financial assistance
public health care insurance
public health care insurance
mentoring
mentoring
documents for employment
documents for employment
place to live
place to live
transportation
transportation
drivers license
drivers license
clothes/food bank
clothes/food bank
taken program to prepare for releaseb
taken class to prepare for releaseb
d
after-school/weekend/summer sports program
after-school/weekend/summer sports programe
Health Service Bundle
medical treatment
medical treatment
dental services
mental health treatment
mental health treatment
substance use treatment
substance use treatment
group for abuse victims
abuse victim support group
anger management services
anger management services
Employment, Education, and Skills Service Bundle
job
any employment services
more education
any educational services
money management skills
money management assistance
life skills
life-skills training
work on personal relationships
help with personal relationships
change attitudes on criminal behavior
training to change attitudes on criminal behavior
Not Applicable
(continued)
46
Service Needs and Service Receipt
Exhibit 13. Items included in service need and service receipt bundles (continued)
Need
Receipt
Domestic Violence–Related Service Bundle
batterer intervention program
batterer intervention program
domestic violence support group
domestic violence support group
Child-Related Service Bundle
child support payments
help getting child support payments
f
modifications in child support debt
help modifying child support debtf
modifications in custody
help modifying custody
parenting skills
parenting skills classes
child care
child care
a
Waves 2 through 4 only.
Wave 1 only.
c
Juveniles only.
d
Juveniles not incarcerated, at Waves 2 through 4 only.
e
Juveniles, at Waves 2 through 4 only.
f
Asked only of those owing back child support.
b
The service needs and
receipt item and bundle
scores were generated by
weighting the
observations with weights
derived from the
propensity scores.
The service needs and receipt item and bundle scores reported
here were generated by weighting the observations with
weights derived from the propensity scores. These scores
provide adjustments for differences in SVORI program
assignment between the SVORI and non-SVORI groups on
observed characteristics. Group means were calculated with
Proc Survey Means (SAS 9.1.3), and significance tests were
conducted through examination of the significance of parameter
estimates on a SVORI indicator variable.
The service needs and services receipt reported by the adult
male respondents are described first, followed by those
reported by the adult female respondents and the juvenile male
respondents. The final section in this chapter assesses
similarities and differences among the three groups.
ADULT MALE SERVICE NEEDS AND RECEIPT
The data suggested high and relatively sustained expressed
need for a variety of services, particularly in the employment/
education/skills and transition services areas. SVORI programs
were successful in greatly increasing access to services, but
reported service receipt was lower than expressed need
throughout.
47
Multi-site Evaluation of SVORI: Summary & Synthesis
Service Needs
The pre-release (Wave 1) data suggested high levels of
expressed need as can be seen in Exhibit 14. The levels of
expressed need for employment, education, and skills were
particularly high—on average, respondents reported needing
nearly three quarters of all of the service items in the
employment bundle (average bundle scores of 74 for SVORI
and non-SVORI). Respondents also expressed a high level of
need for the services and assistance contained in the transition
services bundle. On average, respondents reported needing
nearly two thirds of these services, which include financial
assistance, transportation, and obtaining a driver’s license and
other documentation (average scores of 64 for SVORI and 62
for non-SVORI).
Exhibit 14. Adult male pre-release service need bundle score means across service type, by
group
74
Employment/education/life skills
services
74
64
Transition services
62
46
Child servicesa
48
32
Health services*
34
7
Domestic violence services
Non-SVORI
8
0
a
SVORI
20
40
60
80
100
Among those who reported having minor children.
*p < 0.05 for test of significant difference between SVORI and non-SVORI.
Exhibit 15 provides information for each bundle at each data
collection point for both groups. An overall pattern is apparent
in that expressed need is highest immediately before release,
declines between release and the 3-month follow-up interview,
and then either holds steady or increases slightly over time.
Furthermore, expressed need for employment/education/skills
48
Service Needs and Service Receipt
Exhibit 15. Adult male service need bundle scores for Waves 1 through 4
4
Domestic violence–related services, 4
6
3
3
4
4
7
8
Domestic violence–related services, 3
Domestic violence–related services, 2
Domestic violence–related services, 1
SVORI
Non-SVORI
40
41
38
41
39
39
Child-related services, 4
Child-related services, 3
Child-related services, 2
46
48
Child-related services, 1
28
29
26
27
24
25
32
34
Health services, 4
Health services, 3
Health services, 2
Health services, 1*
49
50
49
49
49
50
Transition services, 4
Transition services, 3
Transition services, 2
64
62
63
64
63
62
60
60
Transition services, 1
Employment/education/skills, 4
Employment/education/skills, 3
Employment/education/skills, 2
74
74
Employment/education/skills, 1
0
20
40
60
80
100
Note: Numbers on labels indicate data collection Waves 1 through 4. Wave 1 = 30 days pre-release; Wave 2 = 3
months post-release; Wave 3 = 9 month post-release; Wave 4 = 15 months post-release.
*p < 0.05 for test of significant difference between SVORI and non-SVORI.
services remains high throughout the follow-up period, with
respondents on average reporting that they needed about two
thirds of the six service items included in this bundle. There
was no significant difference between SVORI and non-SVORI in
expressed need across the service bundles and the data
collection periods.
In addition to domain-specific service bundles, a “super-bundle”
score was calculated that included all service need items. Both
groups had Wave 1 super-bundle scores of 54—indicating an
49
Multi-site Evaluation of SVORI: Summary & Synthesis
Service needs superbundle scores for adult
males
Wave
1
2
3
4
SVORI
NonSVORI
54
42
43
44
54
43
43
45
Note: Differences between
SVORI and non-SVORI
were not significant at the
0.05 level. Wave 1 = 30
days pre-release; Wave 2 =
3 months post-release;
Wave 3 = 9 month postrelease; Wave 4 = 15
months post-release.
More than 85% of adult
male respondents at all
waves reported needing
more education.
expressed need for more than 50% of the service need items
shown in Exhibit 13. Consistent with the component bundles,
the super-bundle scores declined in the three post-release data
collections, as shown in the inset. Average super-bundle scores
for SVORI and non-SVORI groups were 42 and 43 at Wave 2,
43 and 44 at Wave 3, and 44 and 45 at Wave 4.
Before release, the most common service item needs reported
were education (94%), financial assistance (86%), a driver’s
license (83%), job training (82%), and employment (80%).
SVORI and non-SVORI respondents were similar on most
measures, but non-SVORI respondents were significantly less
likely than SVORI respondents to report needing financial
assistance or access to clothing and food banks and more likely
than SVORI respondents to report needing mental health or
substance use treatment, domestic violence support groups, or
a change in their criminal attitudes.15
More than 85% of adult male respondents at all waves reported
needing more education, the highest need of six employment/
education/skills service items. Financial assistance,
transportation, and a driver’s license were the most commonly
reported of 10 transitional service needs pre-release and at 3, 9
and 15 months post-release. Public health care insurance and
financial assistance were also consistently reported as needs by
majorities of both groups.
Service Receipt
SVORI programs were successful in greatly increasing access to
a wide range of services and programming. The SVORI
respondents were much more likely than the non-SVORI
respondents to report receiving most of the services asked
about.
Exhibit 16 shows the service receipt bundle scores for the six
bundles (five need bundles plus coordination) for both groups
and all four data collection waves. As can be seen, SVORI
program participants reported receiving significantly more
15
50
Readers interested in more detail are referred to Lattimore, Visher
and Steffey (2008) and Lattimore, Steffey, and Visher (2009).
Service Needs and Service Receipt
Exhibit 16. Adult male service receipt bundle scores across waves by group
1
2
1
1
1
1
Domestic violence–related services, 4*
Domestic violence–related services, 3
Domestic violence–related services, 2
Domestic violence–related services, 1*
SVORI
Non-SVORI
5
7
9
8
Child-related services, 4
Child-related services, 3
10
Child-related services, 2*
13
18
12
Child-related services, 1*
34
22
11
12
13
13
13
11
Health services, 4
Health services, 3
Health services, 2
Health services, 1*
31
Transition services, 4
6
6
Transition services, 3*
7
Transition services, 2*
9
9
13
Transition services, 1*
28
17
Employment/education/skills, 4*
6
35
8
Employment/education/skills, 3*
8
Employment/education/skills, 2*
8
11
14
Employment/education/skills, 1*
24
Coordination services, 4*
24
Coordination services, 3*
38
30
43
31
Coordination services, 2*
40
Coordination services, 1*
33
0
20
40
57
60
60
80
100
Note: Numbers on labels indicate data collection Waves 1 through 4. Wave 1 = 30 days pre-release; Wave 2 = 3
months post-release; Wave 3 = 9 month post-release; Wave 4 = 15 months post-release.
*p < 0.05 for test of significant difference between SVORI and non-SVORI.
coordination, employment/education/skills, and transition
services than comparison respondents at all interviews. Overall,
levels of reported service receipt declined substantially between
the pre-release and the first post-release interview, and the
differences between SVORI and non-SVORI groups diminished
over time. Aggregate levels of service receipt were substantially
lower than comparable measures of service need (across all
51
Multi-site Evaluation of SVORI: Summary & Synthesis
Service receipt superbundle scores for adult
males
Wave
SVORI
NonSVORI
1*
2*
3*
4
34
18
13
9
22
12
10
8
Note: Wave 1 = 30 days prerelease; Wave 2 = 3
months post-release; Wave
3 = 9 month post-release;
Wave 4 = 15 months postrelease.
*p < 0.05 for test of significant
difference between SVORI
and non-SVORI.
The percentages of
SVORI program
participants who reported
receiving any
employment-related
services was 37% prerelease—a small
proportion that declined
to 34% at 3 months, 21%
at 9 months, and 14% at
15 months post-release
bundles and time periods and among both groups), indicating
that most men had unmet needs. Additionally, although postrelease expressed needs remained fairly constant (or even
increased over time), service receipt steadily declined in all
areas. Overall, pre-release, SVORI respondents reported
receiving about one third of the service items—in contrast to
the one fifth of items that non-SVORI respondents reported
receiving (super-bundle scores of 34 and 22). The average
super-bundle scores declined at Waves 2, 3, and 4, as shown in
the inset.
The most common services SVORI respondents reported
receiving before release were participating in programs to
prepare for release, meeting with a case manager, working with
someone to plan for release, taking a class specifically for
release, and receiving a needs assessment. There were only
four services for which the difference in service receipt between
SVORI and non-SVORI respondents was not significant:
assistance modifying custody agreements, batterer intervention
programs, medical treatment, and assistance accessing public
financial assistance.
The percentages of SVORI program participants who reported
receiving any employment-related services was 37% prerelease—a small proportion that declined to 34% at 3 months,
21% at 9 months, and 14% at 15 months post-release.
Although far less than 100%, these proportions were
significantly higher at all waves than those reported by the
non-SVORI group—20% pre-release, declining to 10% at 15
months.
The services that men were most likely to report receiving after
release were similar across the post-release waves and included
post-release supervision, case management, and needs
assessments.
ADULT FEMALE SERVICE NEEDS AND
RECEIPT
The adult females reported higher expressed need for a variety
of services than the adult men across all data waves. The
women also reported higher service receipt than the men, with
women who participated in SVORI programs reporting higher
levels of services than non-SVORI comparison respondents. As
with the men, although SVORI programs were successful in
52
Service Needs and Service Receipt
greatly increasing access to services, reported service receipt
by the women was lower than expressed need throughout.
Service Needs
Service need was
significantly higher for
women than it was for
men across several
service areas (primarily
health services and
family services), at each
time period.
Exhibit 17. Adult female
pre-release service need
bundle scores across
service type, by group
The pre-release (Wave 1) data suggested high levels of
expressed need for these incarcerated women and little
difference between the two groups. Exhibit 17 shows that the
women expressed need for most items in the
employment/education/skills bundle and the transition services
bundle. The levels of expressed need for employment,
education, and skills were particularly high—on average,
respondents reported needing nearly 80% of all the service
items in the employment bundle (average bundle scores of 80
for SVORI and 79 for non-SVORI), and 95% of both groups said
that they needed more education. Respondents also expressed
a high level of need for the services and assistance contained in
the transition services bundle. On average, respondents
reported needing nearly two thirds of these services, which
include financial assistance, transportation, and obtaining a
driver’s license and other documentation (average scores of 74
for SVORI and 72 for non-SVORI).
80
79
Employment/
education/skills, 1
74
72
Transition services, 1
52
48
Child-related services, 1
55
57
Health services, 1
21
21
Domestic violence–
related services, 1
0
20
SVORI
Non-SVORI
40
60
80
100
Note: Numbers on labels indicate data collection Wave 1.
*p < 0.05 for test of significant difference between SVORI and non-SVORI.
Nearly 80% of the women
reported needing medical
treatment.
Expressed need for health services before release was also
high, with respondents on average reporting needing more than
two of the five health services. In particular, nearly 80%
reported needing medical treatment, between 55% and 56%
reported needing mental health treatment, and about two
53
Multi-site Evaluation of SVORI: Summary & Synthesis
thirds reported needing substance use treatment. Overall,
among the needs the women most commonly reported were
education (95%), public health insurance (91%), financial
assistance (87%), employment (83%), and a mentor (83%).
Service needs superbundle scores for adult
females
Wave
1
2
3
4*
SVORI
NonSVORI
65
51
46
43
64
51
47
50
Note: Wave 1 = 30 days prerelease; Wave 2 = 3
months post-release; Wave
3 = 9 month post-release;
Wave 4 = 15 months postrelease.
*p < 0.05 for test of significant
difference between SVORI
and non-SVORI.
Compared with the extremely high self-reported service needs
at the pre-release interview, women reported needing
substantially fewer services at the 3-month post-release
interview. Exhibit 17 provides information for each bundle at
each data collection point for both groups. An overall pattern is
apparent in that expressed need is highest immediately before
release, declines between release and the 3-month follow-up
interview, and then either holds steady or increases slightly
over time. Nonetheless, absolute levels of service need
remained quite high. Women continued to report high levels of
service need (in the 40–50% range) for many services even 15
months after release.
The needs most commonly reported by the women across all 3
follow-up waves were more education (87–93%), public health
care insurance (66–77%), and financial assistance (64–73%).
Furthermore, expressed need for employment/education/skills
services remained high throughout the follow-up period, with
respondents on average reporting that they needed about two
thirds of the six service items included in this bundle
throughout.
As a summary of service need, the super-bundle scores, which
captured the level of overall need across all services, were
lower for the follow-up period than for the pre-release period.
On average, the women reported needing 46–51% of all the
service items during the post-release period, compared with
64% of items during the pre-release period.
Except for the SVORI respondents’ slightly lower pre-release
need for health services as compared with the non-SVORI
respondents’ need (scores of 31 and 34, respectively), there
was no significant difference between SVORI and non-SVORI in
expressed need across the service bundles and the data
collection periods.
54
Service Needs and Service Receipt
Exhibit 18. Adult female service need bundle scores for Waves 1 through 4
10
11
12
12
Domestic violence–related services, 4
Domestic violence–related services, 3
SVORI
Non-SVORI
8
8
Domestic violence–related services, 2
21
21
Domestic violence–related services, 1
35
Child-related services, 4
Child-related services, 3
Child-related services, 2
43
38
39
41
Child-related services, 1
46
48
52
35
Health services, 4
39
37
38
39
41
Health services, 3
Health services, 2
55
57
Health services, 1
48
Transition services, 4*
56
51
52
Transition services, 3
60
60
Transition services, 2
74
72
Transition services, 1
57
Employment/education/skills, 4
64
60
60
Employment/education/skills, 3
64
64
Employment/education/skills, 2
80
79
Employment/education/skills, 1
0
20
40
60
80
100
Note: Numbers on labels indicate data collection Waves 1 through 4. Wave 1 = 30 days pre-release; Wave 2 = 3
months post-release; Wave 3 = 9 month post-release; Wave 4 = 15 months post-release.
*p < 0.05 for test of significant difference between SVORI and non-SVORI.
55
Multi-site Evaluation of SVORI: Summary & Synthesis
Service Receipt
Service receipt superbundle scores for adult
females
Wave
SVORI
NonSVORI
1*
2*
3*
4*
45
28
21
16
24
13
12
12
Note: Wave 1 = 30 days prerelease; Wave 2 = 3
months post-release; Wave
3 = 9 month post-release;
Wave 4 = 15 months postrelease.
*p < 0.05 for test of significant
difference between SVORI
and non-SVORI.
Women who
participated in SVORI
programming were
more likely to receive
programming and
services, which
indicates that SVORI
funding did increase
access to services for
female prisoners
returning to society.
56
Women in SVORI programs had greater access than women
assigned to “treatment as usual” to a wide range of pre-release
services and were more likely to receive most of the services
documented. Exhibit 19 shows the service bundle scores by
group and data collection wave. The women enrolled in SVORI
reported significantly higher levels of service receipt before
release than the non-SVORI respondents across 22 of the 36
services. The most common pre-release services SVORI
respondents reported receiving were participating in programs
to prepare for release, taking a class specifically for release,
working with someone to plan for release, receiving a needs
assessment, and developing a reentry plan. Overall, prerelease, SVORI respondents reported receiving about half of the
service items—in contrast to the one quarter that non-SVORI
respondents reported receiving.
As with the decline of self-reported need for services over time,
the likelihood of receiving services declined over time—although
the decline in service receipt was steeper. Aggregate levels of
service receipt were substantially lower than comparable
measures of service need (across all bundles and time periods
and among both groups), indicating that very small proportions
of women received the services they needed. Over time, the
number of differences in service receipt between the SVORI and
non-SVORI groups decreased at each follow-up wave.
The services that women were most likely to receive after
release were similar across the post-release waves and included
post-release supervision, case management, and needs
assessments.
Service Needs and Service Receipt
Exhibit 19. Adult female service receipt bundle scores across waves, by group
Domestic violence–related services, 4
6
2
Domestic violence–related services, 3
1
Domestic violence–related services, 2*
1
4
SVORI
Non-SVORI
3
Domestic violence–related services, 1*
15
8
Child-related services, 4
Child-related services, 3*
1
6
5
5
6
5
Child-related services, 2
Child-related services, 1*
19
10
22
19
Transition services, 4
Transition services, 3*
18
Transition services, 2*
14
26
25
Transition services, 1*
48
36
13
12
Transition services, 4
Transition services, 3*
17
10
Transition services, 2*
26
12
Transition services, 1*
40
23
Employment/education/skills, 4*
8
Employment/education/skills, 3*
8
Employment/education/skills, 2*
7
14
17
26
Employment/education/skills, 1*
25
Coordination services, 4*
25
Coordination services, 3*
54
39
55
30
Coordination services, 2*
69
36
Coordination services, 1*
73
29
0
20
40
60
80
100
Note: Numbers on labels indicate data collection Waves 1 through 4. Wave 1 = 30 days pre-release; Wave 2 = 3
months post-release; Wave 3 = 9 month post-release; Wave 4 = 15 months post-release.
*p < 0.05 for test of significant difference between SVORI and non-SVORI.
JUVENILE MALE SERVICE NEEDS AND
RECEIPT
The patterns for service needs expressed by the juvenile males
were similar to those for the adults described earlier—relatively
high levels of expressed need pre-release, particularly for
education and employment services, and declining levels in the
post-release interviews, with few differences between the
SVORI and non-SVORI groups. However, unlike the results for
57
Multi-site Evaluation of SVORI: Summary & Synthesis
adult respondents, those for juvenile respondents showed few
differences between the SVORI and non-SVORI groups in
reports of service receipt—whether before or after release from
detention.
Service Needs
Nearly all (95%) of the
juvenile male respondents
reported needing more
education, and between
85% and 90% reported
needing job training and
a job.
Exhibit 20. Juvenile
male pre-release service
need bundle scores
across service type, by
group
Exhibit 20 shows the weighted pre-release service need bundle
scores for the juvenile male respondents. As can be seen, on
average, the juvenile males indicated needing about three
fourths of the seven employment/education/skills services.
Within this bundle, nearly all (95%) of the juvenile male
respondents reported needing more education, and between 85%
and 90% reported needing job training and a job (data not
shown). The respondents also reported needing about half of the
transition services, with a driver’s license (about 90%),
transportation (about two thirds), and a mentor (about 60%)
being the most common of the 10 transition services expressed
as needs. Somewhat less than a third of the 5 health services
were identified by the juvenile males as needs—about half
reported needing anger management and medical treatment, the
most commonly identified needs among the health service items.
75
75
Employment/education/skills, 1
52
49
Transition services, 1
29
27
Child-related services, 1
32
30
Health services, 1
SVORI
7
9
Domestic violence–related services, 1
0
Non-SVORI
20
40
60
80
Note: Numbers on labels indicate data collection Wave 1.
*p < 0.05 for test of significant difference between SVORI and non-SVORI.
58
100
Service Needs and Service Receipt
Service needs superbundle scores for
juvenile males
Wave
1
2
3
4
SVORI
NonSVORI
49
36
38
35
49
35
35
38
Note: Differences between
SVORI and non-SVORI
were not significant at the
0.05 level. Wave 1 = 30
days pre-release; Wave 2 =
3 months post-release;
Wave 3 = 9 month postrelease; Wave 4 = 15
months post-release.
About 9% of the juvenile males (N = 30) reported having one
or more children. These individuals reported needing less than
a third of the five child-related services (bundle scores of 29 for
SVORI and 27 for non-SVORI), with parenting skills classes
being the most commonly expressed. Few of these young
males expressed a need for either domestic violence programs
or batterer intervention classes.
Before their release from confinement, respondents reported
needing, on average, slightly less than half of the wide array of
services measured; super-bundle service need scores were 49
for both groups. Overall, before release from confinement, the
most common needs reported were more education (95%), a
driver’s license (90%), job training (88%), a job (87%), and
life-skills training (76%).
Exhibit 21 shows the bundle scores for both groups at all data
collection waves. As can be seen, scores were highest for
employment/education/skills services at all waves—declining from
an average report of needing 75% of the seven services at wave 1
to about 60% after release. Overall, respondents reported levels
of service need that were lower than their pre-release levels of
need. At each post-release interview, respondents reported that
they needed, on average, more than one third of the services
measured (super-bundle scores ranged from 35 to 38).
In the post-release interviews, at least half of SVORI
respondents reported that they needed more education, a
driver’s license, a job, job training, transportation, and lifeskills training (data not shown). Similar levels of need were
reported by non-SVORI respondents. At each post-release
interview, SVORI and non-SVORI respondents were similar on
most service need measures; however, 9 months after release,
SVORI respondents were significantly more likely than nonSVORI counterparts to report that they needed life-skills
training; 15 months after release, non-SVORI respondents were
significantly more likely than SVORI respondents to report that
they needed transportation.
59
Multi-site Evaluation of SVORI: Summary & Synthesis
Exhibit 21. Juvenile male service need bundle scores for Waves 1 through 4
Domestic violence–related services, 4
3
Domestic violence–related services, 3
3
Domestic violence–related services, 2
5
6
Non-SVORI
9
4
7
Domestic violence–related services, 1
SVORI
9
24
Child-related services, 4
Child-related services, 3
35
50
26
19
Child-related services, 2
41
29
27
Child-related services, 1
Health services, 4
17
Health services, 3
17
Health services, 2
35
37
37
18
32
30
Health services, 1
37
Transition services, 4
42
40
41
38
39
Transition services, 3
Transition services, 2
Transition services, 1
49
52
58
59
Employment/education/skills, 4
Employment/education/skills, 3*
54
56
52
Employment/education/skills, 2
63
75
75
Employment/education/skills, 1
0
20
40
60
80
100
Note: Numbers on labels indicate data collection Waves 1 through 4. Wave 1 = 30 days pre-release; Wave 2 = 3
months post-release; Wave 3 = 9 month post-release; Wave 4 = 15 months post-release..
*p < 0.05 for test of significant difference between SVORI and non-SVORI.
60
Service Needs and Service Receipt
Service Receipt
Service receipt superbundle scores for
juvenile males
Wave
SVORI
NonSVORI
1
2*
3
4
38
19
12
9
37
16
10
7
Note: Wave 1 = 30 days prerelease; Wave 2 = 3
months post-release; Wave
3 = 9 month post-release;
Wave 4 = 15 months postrelease.
*p < 0.05 for test of significant
difference between SVORI
and non-SVORI.
SVORI programs achieved only modest increases in providing a
wide range of pre-release services and programs. Overall,
SVORI respondents were more likely than non-SVORI
respondents to report receiving most of the 60 services
measured; however, unlike differences observed for the adult
programs, few of these differences were statistically significant.
Exhibit 22 shows the service receipt bundle scores for the four
data collection periods. Overall, the reported levels of service
receipt were highest for SVORI and non-SVORI respondents
before their release from confinement, declined dramatically in
the 3 months after release, and remained low throughout the
post-release period. Before release, respondents reported
receiving less than 40% of all services and less than 20% after
release (see sidebar). The most common services SVORI
respondents reported having received before their release from
confinement were educational services (94%), a meeting with a
case manager (90%), a needs assessment (83%), collaboration
with someone to plan for release (78%), and medical treatment
(73%).
Although SVORI and non-SVORI respondents reported low
levels of post-release service receipt, SVORI respondents
generally reported higher levels of service receipt than nonSVORI respondents. In fact, 3 months after release, SVORI
respondents reported receiving significantly more health and
employment/education/skills services than non-SVORI
respondents. The most common post-release services SVORI
respondents reported receiving were a meeting with a case
manager, a needs assessment, educational services,
collaboration with someone to reintegrate into the community,
and employment services. At each post-release period and for
each service bundle, the levels of service receipt reported by
SVORI and non-SVORI respondents were considerably lower
than their reported levels of service need.
61
Multi-site Evaluation of SVORI: Summary & Synthesis
Exhibit 22. Juvenile male service receipt bundle scores across waves, by group
0
0
0
0
0
1
Domestic violence–related services, 4
Domestic violence–related services, 3
Domestic violence–related services, 2
SVORI
Non-SVORI
Domestic violence–related services, 1
5
Child-related services, 4
1
Child-related services, 3
3
2
0
Child-related services, 2
Child-related services, 1
7
4
2
17
6
Health services, 4
Health services, 3
7
8
Health services, 2*
7
10
10
12
42
Health services, 1*
Transition services, 4*
5
47
8
Transition services, 3
8
7
Transition services, 2
8
12
23
21
Transition services, 1
Employment/education/skills, 4
9
Employment/education/skills, 3
14
18
13
Employment/education/skills, 2*
16
21
52
50
Employment/education/skills, 1
13
14
Coordination services, 4
Coordination services, 3*
18
24
43
42
Coordination services, 2
Coordination services, 1
53
0
20
40
57
60
80
100
Note: Numbers on labels indicate data collection Waves 1 through 4. Wave 1 = 30 days pre-release; Wave 2 = 3
months post-release; Wave 3 = 9 month post-release; Wave 4 = 15 months post-release.
*p < 0.05 for test of significant difference between SVORI and non-SVORI.
62
Service Needs and Service Receipt
COMPARISON OF ADULT MALE, ADULT
FEMALE, AND JUVENILE MALE SVORI
EVALUATION PARTICIPANTS
In this section, comparisons among the three demographic
groups are presented, focusing primarily on the transition,
employment/education/skills, health, and, for service receipt,
coordination service bundles. Needs are discussed first and
then receipt.
Service Needs
Exhibit 23 shows the weighted service need super-bundle
scores for the adult males, adult females, and juvenile males by
group and data collection wave. As can be seen, before and
immediately after release, the expressed needs of the women
are significantly larger than the needs expressed by the men
and juvenile males. Additionally, the juvenile males reported
lower numbers of needs than the adults across all data
collection points, although it is worth remembering that the
overall needs of the juvenile males are somewhat suppressed
by low demand for child-related services and domestic
violence–related services, which comprise a total of seven
service items.
Exhibit 24 shows the employment/education/skills bundle
scores for the adult men, adult women, and juvenile males by
SVORI and non-SVORI study group and four data collection
waves. As noted previously, most respondents reported high
needs for most of the services included in this service bundle
(job training, job, education, money management skills, life
skills, help working on personal relationships, change in
criminal behavior attitudes)—particularly before release. In
addition to similar reports by SVORI and non-SVORI groups,
the three demographic groups reported similar levels of needs
at each wave.16
16
The differences between the SVORI and non-SVORI means for the
women at Wave 2 and the juvenile males at Wave 3 were
statistically significant at the alpha = 0.05 level.
63
Multi-site Evaluation of SVORI: Summary & Synthesis
Exhibit 23. Weighted service need super-bundle scores by demographic group, wave, and
study group
100
90
80
70
60
50
40
30
20
10
0
65
54
64
54
49
49
100
90
80
70
60
50
40
30
20
10
0
51
42
36
100
90
80
70
60
50
40
30
20
10
0
43
43
Men, non-SVORI
Women, non-SVORI
Juvenile, non-SVORI
46
35
Wave 2
Wave 1
Men, SVORI
Women, SVORI
Juvenile, SVORI
51
43
47
38
35
Men, SVORI
Women, SVORI
Juvenile, SVORI
100
90
80
70
60
50
40
30
20
10
0
44
45
Men, non-SVORI
Women, SVORI
Women, non-SVORI
Juvenile, SVORI
Juvenile, non-SVORI
50
43
35
38
Wave 4
Wave 3
Men, SVORI
Men, non-SVORI
Women, non-SVORI
Juvenile, non-SVORI
Men, SVORI
Women, SVORI
Juvenile, SVORI
Men, non-SVORI
Women, non-SVORI
Juvenile, non-SVORI
Note: Wave 1 = 30 days pre-release; Wave 2 = 3 months post-release; Wave 3 = 9 month post-release; Wave 4 = 15
months post-release.
64
Service Needs and Service Receipt
Exhibit 24. Employment/education/skills need bundle scores by demographic group, study
group, and data collection wave
100
90
80
70
60
50
40
30
20
10
0
74
80
74
79
75
75
100
90
80
70
60
50
40
30
20
10
0
60
60
Wave 1
Men, SVORI
Women, SVORI
Juvenile, SVORI
100
90
80
70
60
50
40
30
20
10
0
63
62
60
56
63
54
Men, SVORI
Women, SVORI
Juvenile, SVORI
100
90
80
70
60
50
40
30
20
10
0
63
52
Men, non-SVORI
Women, non-SVORI
Juvenile, non-SVORI
Men, non-SVORI
Women, non-SVORI
Juvenile, non-SVORI
64
64
57
58
59
Wave 4
Wave 3
Men, SVORI
Women, SVORI
Juvenile, SVORI
64
Wave 2
Men, non-SVORI
Women, non-SVORI
Juvenile, non-SVORI
60
64
Men, SVORI
Women, SVORI
Juvenile, SVORI
Men, non-SVORI
Women, non-SVORI
Juvenile, non-SVORI
Note: Wave 1 = 30 days pre-release; Wave 2 = 3 months post-release; Wave 3 = 9 month post-release; Wave 4 = 15
months post-release.
The levels of expressed need for employment/education/skills
services decreased after release and remained relatively
constant over the three post-release data collection periods.
The need for more education dominated this service need
bundle—with most respondents at all waves indicating that they
65
Multi-site Evaluation of SVORI: Summary & Synthesis
needed more education. Specifically, at least 85% of each
group reported needing more education throughout (data not
shown).
Exhibit 25 shows the expressed needs for the multiple services
in the transition services bundle that included a variety of
items, such as a driver’s license, identification for employment,
transportation, housing, and health care. The ordering among
the groups remains similar across time. The women have the
highest expressed need for transition services, followed by the
men, and then the juvenile males. The expressed need declines
after release, with respondents indicating that they need, on
average, about half of the transition service items. As noted
earlier, most respondents pre-release reported needing help
obtaining a driver’s license. The juvenile males were less likely
to report needing public health care insurance—only about 50%
of the juveniles, in comparison with more than 90% of the
adult females and about 75% of the adult males who indicated
that they needed public health care insurance (data not
shown). The juveniles were also less likely to indicate that they
needed help finding a place to live—about one quarter of the
juveniles, in contrast to about one half of the adults who
indicated before release that they needed a place to live.
Exhibit 26 shows the data for the health services bundle scores.
As can be seen, the women expressed much higher needs for
health-related services than the men or juvenile males. The
health needs bundle included five items—medical treatment,
mental health treatment, substance use treatment, abuse
victim group, and anger management. The women’s bundle
scores suggest that that they needed about two or more of
these services—in contrast to the males, who needed 30% or
less of the five items, or one to two items, on average. Overall,
all groups were most likely to indicate a need for medical
treatment, but there were substantial differences among the
groups. Nearly 80% of the women, in comparison to less than
60% of the men and about 45% of the juvenile males,
expressed a need for medical treatment before release (data
not shown); these proportions declined after release so that at
15 months post-release about 60% of the women, 55% of the
men, and less than 40% of the juvenile males reported needing
medical treatment.
66
Service Needs and Service Receipt
Exhibit 25. Transition services need bundle scores by demographic group, study group, and
data collection wave
100
90
80
70
60
50
40
30
20
10
0
74
64
72
62
52
49
100
90
80
70
60
50
40
30
20
10
0
60
49
38
Wave 1
Men, SVORI
Women, SVORI
Juvenile, SVORI
100
90
80
70
60
50
40
30
20
10
0
49
49
52
40
41
Men, SVORI
Women, SVORI
Juvenile, SVORI
100
90
80
70
60
50
40
30
20
10
0
49
50
Men, non-SVORI
Women, non-SVORI
Juvenile, non-SVORI
Men, non-SVORI
Women, non-SVORI
Juvenile, non-SVORI
56
48
37
42
Wave 4
Wave 3
Men, SVORI
Women, SVORI
Juvenile, SVORI
39
Wave 2
Men, non-SVORI
Women, non-SVORI
Juvenile, non-SVORI
51
60
50
Men, SVORI
Women, SVORI
Juvenile, SVORI
Men, non-SVORI
Women, non-SVORI
Juvenile, non-SVORI
Note: Wave 1 = 30 days pre-release; Wave 2 = 3 months post-release; Wave 3 = 9 month post-release; Wave 4 = 15
months post-release.
67
Multi-site Evaluation of SVORI: Summary & Synthesis
Exhibit 26. Health needs bundle scores by demographic group, study group, and data
collection wave
100
90
80
70
60
50
40
30
20
10
0
55
32
57
34
30
32
100
90
80
70
60
50
40
30
20
10
0
39
24
25
18
100
90
80
70
60
50
40
30
20
10
0
Men, non-SVORI
Women, non-SVORI
Juvenile, non-SVORI
37
26
38
27
17
15
Men, SVORI
Women, SVORI
Juvenile, SVORI
100
90
80
70
60
50
40
30
20
10
0
Men, non-SVORI
Women, non-SVORI
Juvenile, non-SVORI
Men, non-SVORI
Women, non-SVORI
Juvenile, non-SVORI
35
28
39
29
17
18
Wave 4
Wave 3
Men, SVORI
Women, SVORI
Juvenile, SVORI
15
Wave 2
Wave 1
Men, SVORI
Women, SVORI
Juvenile, SVORI
41
Men, SVORI
Women, SVORI
Juvenile, SVORI
Men, non-SVORI
Women, non-SVORI
Juvenile, non-SVORI
Note: Wave 1 = 30 days pre-release; Wave 2 = 3 months post-release; Wave 3 = 9 month post-release; Wave 4 = 15
months post-release.
68
Service Needs and Service Receipt
Relatively large percentages of the women reported needing
mental health and substance use treatment. More than half
reported needing mental health treatment before release, a
proportion that declined to about 40% across the post-release
interviews. About two thirds reported needing substance use
treatment before release, while 40% or less reported needing
substance use treatment after release (data not shown). Many
fewer men and juvenile males reported needing either mental
health or substance use treatment.
Service Receipt
Exhibit 27 shows the weighted super-bundle scores for service
receipt by group and wave. On average, women who were
participating in SVORI programs reported receiving the largest
proportion of services across all data collection waves.
Interestingly, the levels of services reported by the non-SVORI
women and men were quite similar—suggesting that the status
quo in these states was similar for adult males and for adult
females. The rapid decline in services after release is also
apparent for all groups. Before release, respondents reported
receiving between 22% (non-SVORI men) and 45% (SVORI
women) of service items. At 15 months post-release, the
percentage had declined to 16% or less for all groups, with
currently being on supervision or other factors included in the
coordination bundle being the most likely “item” reported by all
groups.
Coordination services included receiving a needs assessment,
meeting with a case manager, working with someone to plan
for release (pre-release) or to reintegrate (post-release), being
currently on probation/parole (post-release), developing a
reentry plan (pre-release), and receiving a needs assessment
specifically for release (pre-release). The coordination services
bundle scores are shown in Exhibit 28. The women participating
in SVORI programs reported the greatest receipt of services in
this bundle across all waves of data collection—followed by the
adult men who were in SVORI programs. The levels of services
for the non-SVORI comparison men and women are very
similar across the waves.
69
Multi-site Evaluation of SVORI: Summary & Synthesis
Exhibit 27. Service receipt super-bundle scores by demographic group, study group, and
data collection wave
100
90
80
70
60
50
40
30
20
10
0
45
38
34
37
24
22
100
90
80
70
60
50
40
30
20
10
0
28
18
100
90
80
70
60
50
40
30
20
10
0
Men, non-SVORI
Women, non-SVORI
Juvenile, non-SVORI
21
13
10
12
19
16
Wave 2
Wave 1
Men, SVORI
Women, SVORI
Juvenile, SVORI
13
12
12
10
Men, SVORI
Women, SVORI
Juvenile, SVORI
100
90
80
70
60
50
40
30
20
10
0
16
9
8
Wave 3
Men, SVORI
Men, non-SVORI
Women, SVORI
Women, non-SVORI
Juvenile, SVORI
Juvenile, non-SVORI
Men, non-SVORI
Women, non-SVORI
Juvenile, non-SVORI
12
9
7
Wave 4
Men, SVORI
Women, SVORI
Juvenile, SVORI
Men, non-SVORI
Women, non-SVORI
Juvenile, non-SVORI
Note: Wave 1 = 30 days pre-release; Wave 2 = 3 months post-release; Wave 3 = 9 month post-release; Wave 4 = 15
months post-release.
70
Service Needs and Service Receipt
Exhibit 28. Coordination service receipt bundle scores by demographic group, study group,
and data collection wave
100
90
80
70
60
50
40
30
20
10
0
73
60
57
33
53
29
100
90
80
70
60
50
40
30
20
10
0
69
57
Wave 1
Men, SVORI
Women, SVORI
Juvenile, SVORI
100
90
80
70
60
50
40
30
20
10
0
Men, non-SVORI
Women, non-SVORI
Juvenile, non-SVORI
43
30
24
18
Men, SVORI
Women, SVORI
Juvenile, SVORI
100
90
80
70
60
50
40
30
20
10
0
42
Men, non-SVORI
Women, non-SVORI
Juvenile, non-SVORI
Men, non-SVORI
Women, non-SVORI
Juvenile, non-SVORI
39
30
24
Wave 3
Men, SVORI
Women, SVORI
Juvenile, SVORI
36
Wave 2
55
31
43
40
25
13
14
Wave 4
Men, SVORI
Women, SVORI
Juvenile, SVORI
Men, non-SVORI
Women, non-SVORI
Juvenile, non-SVORI
Note: Wave 1 = 30 days pre-release; Wave 2 = 3 months post-release; Wave 3 = 9 month post-release; Wave 4 = 15
months post-release.
The coordination services bundle for the juvenile male includes
two additional items—assistance accessing a child welfare
caseworker and meeting with a child welfare caseworker. These
services were not reported by many of the juvenile respondents
post-release and, therefore, the many negative responses to
these two items effectively “diluted” the coordination service
71
Multi-site Evaluation of SVORI: Summary & Synthesis
bundle scores.17 However, a substantial decline also occurred
between Wave 2 and Wave 4 in the proportion of the juveniles
who reported being on probation/parole. Three months after
release 90% of SVORI program participants and 82% of nonSVORI comparison respondents reported being on supervision.
Nine months after release the percentage declined to 45% of
both groups, and at 15 months it declined to 39% of the SVORI
and 28% of the non-SVORI groups. Concurrent declines were
observed in the other key categories in the bundle—receiving a
needs assessment, meeting with someone on reintegration, and
meeting with a case manager.18
Exhibit 29 presents the employment/education/skills service
receipt bundle scores. Again, the adult female SVORI program
participants reported more services, although the level of
services reported by the juvenile males was similar. The decline
after release in the proportion of the six services reported as
received is striking for all groups—declining at least 50% across
all groups. Overall, it seems that these individuals received
little assistance with respect to employment, education, and
skill needs—even though most respondents expressed very
high need for help, particularly with respect to education and
employment (see Exhibit 24). Among the SVORI participants,
employment services were the most likely to be reported after
release.
17
18
72
There are four items in the adult coordination bundle and six in the
juvenile male bundle. The proportions of juveniles reporting
assistance accessing child welfare caseworker was about 9% at
Wave 2, less than 5% at Wave 3, and about 2% at Wave 4.
Similarly, only about 17% of juveniles reported meeting with a child
welfare caseworker at Wave 2; 7%, at Wave 3; and less than 5%,
at Wave 4.
The coordination service receipt bundle scores were calculated with
the adult scoring algorithm for the juvenile males’ data (i.e.,
excluding the two questions about child welfare caseworker). As
expected, the bundle scores were higher, particularly for Wave 1
and 2 data. Specifically, scores were 68 and 62 for SVORI and nonSVORI at Wave 1 (compare 57 and 53 in Exhibit 28), 58 and 56 at
Wave 2 (compare 43 and 42), 34 and 25 at Wave 3 (compare 24
and 18), and 19 and 20 at Wave 4 (compare 13 and 14).
Service Needs and Service Receipt
Exhibit 29. Employment/education/skills service receipt bundle scores by demographic
group, study group, and data collection wave
100
90
80
70
60
50
40
30
20
10
0
54
52
50
38
25
24
100
90
80
70
60
50
40
30
20
10
0
26
14
8
Wave 1
Men, SVORI
Women, SVORI
Juvenile, SVORI
100
90
80
70
60
50
40
30
20
10
0
11
18
8
13
Men, SVORI
Women, SVORI
Juvenile, SVORI
100
90
80
70
60
50
40
30
20
10
0
8
Men, non-SVORI
Women, non-SVORI
Juvenile, non-SVORI
Men, non-SVORI
Women, non-SVORI
Juvenile, non-SVORI
14
6
8
14
9
Wave 4
Wave 3
Men, SVORI
Women, SVORI
Juvenile, SVORI
16
7
Wave 2
Men, non-SVORI
Women, non-SVORI
Juvenile, non-SVORI
17
8
21
Men, SVORI
Women, SVORI
Juvenile, SVORI
Men, non-SVORI
Women, non-SVORI
Juvenile, non-SVORI
Note: Wave 1 = 30 days pre-release; Wave 2 = 3 months post-release; Wave 3 = 9 month post-release; Wave 4 = 15
months post-release.
73
Multi-site Evaluation of SVORI: Summary & Synthesis
Results for transition services are shown in Exhibit 30. There
were 10 items (12 items at Wave 1) in the adult bundle and an
additional item for after-school/weekend/summer sports for the
juvenile males post-release.19 The respondents in all groups
reported receiving help on some items before release, but the
likelihood of receiving help on any item declined dramatically
after release. As noted previously, adult SVORI program
participants received significantly more transition services than
non-SVORI comparisons through the 9-month post-release
interview. There was little difference between the two juvenile
groups.20
The transition services included services and items directed at
reintegration—help accessing housing, documents for
employment, a driver’s license, transportation, clothing and
food, mentoring, public health care insurance, financial
assistance, public financial assistance, and legal assistance, as
well as programs or classes to prepare for release (Wave 1
only). Exhibit 30 shows that individuals reported receiving few
of these services.
The women who were in SVORI programs reported receiving,
on average, help with about 5 of the 12 items pre-release
(Wave 1); and, of the 10 post-release items, they reported
receiving between 2 and 3 during the 3 months after release
and less than 2 in subsequent months. The most commonly
reported services received before release were programs (89%)
or classes (82%) to prepare for release and help obtaining
documents for employment (59%). Post-release, help accessing
public financial assistance and health care were among the
services most commonly reported by the women.
19
20
74
As described for the coordination bundles, the juvenile males’
transition bundle scores were calculated with the adult scoring
algorithm (i.e., omitting the after-school/weekend/summer sports
program question). In this case, it made little difference in the
scores. SVORI and non-SVORI scores were 13 and 8 (Wave 2), 8
and 7 (Wave 3), and 8 and 5 (Wave 4). The bundles were the same
at Wave 1.
Although the difference in transition service receipt bundle scores at
Wave 4 was statistically significant, the overall levels were so low
(8 for the SVORI participants and 5 for the non-SVORI
comparisons) as to make this difference not meaningful.
Service Needs and Service Receipt
Exhibit 30. Transition receipt service bundle scores by demographic group, study group,
and data collection wave
100
90
80
70
60
50
40
30
20
10
0
40
28
23
17
23
21
100
90
80
70
60
50
40
30
20
10
0
26
13
Wave 1
Men, SVORI
Women, SVORI
Juvenile, SVORI
100
90
80
70
60
50
40
30
20
10
0
7
Men, non-SVORI
Women, non-SVORI
Juvenile, non-SVORI
10
8
7
Men, SVORI
Women, SVORI
Juvenile, SVORI
100
90
80
70
60
50
40
30
20
10
0
8
Men, non-SVORI
Women, non-SVORI
Juvenile, non-SVORI
Men, non-SVORI
Women, non-SVORI
Juvenile, non-SVORI
13
6
6
12
8
5
Wave 4
Wave 3
Men, SVORI
Women, SVORI
Juvenile, SVORI
12
Wave 2
17
9
12
9
Men, SVORI
Women, SVORI
Juvenile, SVORI
Men, non-SVORI
Women, non-SVORI
Juvenile, non-SVORI
Note: Wave 1 = 30 days pre-release; Wave 2 = 3 months post-release; Wave 3 = 9 month post-release; Wave 4 = 15
months post-release.
Before release, the adult male SVORI program participants
were most likely to report receiving programs (75%) and
classes (65%) to prepare for release, as well as assistance
obtaining documents for employment (e.g., Social Security
card). Reports of service receipt declined markedly after
75
Multi-site Evaluation of SVORI: Summary & Synthesis
release, when 25% of SVORI program participants, in the 3
months since release, reported receiving help obtaining
documents and when 20% reported help accessing public
financial assistance; reports for all other items were
substantially lower. At the 9- and 15-month interviews, less
than 15% and 7% reported having received any of these items
in the previous 6 months.
As is evident in Exhibit 30, the juvenile males reported
receiving few transition services—even before release. There
were few differences between the SVORI and SVORI groups,
with both most likely to report receiving programs and classes
to prepare for release and mentoring as the most common prerelease items. Post-release, less than 20% reported receiving
each of the services.
The women who were in
SVORI programs were
significantly more likely
than the non-SVORI
comparison respondents,
at each interview, to
report receipt of mental
health and substance use
treatment.
Exhibit 31 shows the health services receipt bundle scores
(medical treatment, dental treatment, mental health treatment,
substance use treatment, group for abuse victims, and anger
management programs). Because prisons and detention centers
are required by law to provide medical and dental services,
there should not have been differences in the proportions
reporting receiving these services pre-release and, overall,
there were not; therefore, differences between the SVORI and
non-SVORI groups, if any, were driven by differences in receipt
of the other services.
Exhibit 31 presents a familiar pattern—more services for all
groups before release, with women reporting more services
than the adult and juvenile males (at Wave 1, both groups of
juvenile males reported higher health services than the nonSVORI women). The women who were in SVORI programs were
significantly more likely than the non-SVORI comparison
respondents to report receiving mental health and substance
use treatment at each interview. Men who were in SVORI
programs were more likely than non-SVORI comparisons to
report participating in anger management programs and
substance use treatment before release, with few other
differences being evident. Substantial proportions of the
juvenile males reported receiving anger management and
substance use treatment before release; few reported receiving
any health services post-release.
76
Service Needs and Service Receipt
Exhibit 31. Health services receipt bundle scores by demographic group, study group, and
data collection wave
100
90
80
70
60
50
40
30
20
10
0
48
35
42
47
36
31
100
90
80
70
60
50
40
30
20
10
0
25
13
Wave 1
Men, SVORI
Women, SVORI
Juvenile, SVORI
100
90
80
70
60
50
40
30
20
10
0
13
Men, non-SVORI
Women, non-SVORI
Juvenile, non-SVORI
18
8
10
Men, SVORI
Women, SVORI
Juvenile, SVORI
100
90
80
70
60
50
40
30
20
10
0
7
Men, non-SVORI
Women, non-SVORI
Juvenile, non-SVORI
Men, non-SVORI
Women, non-SVORI
Juvenile, non-SVORI
22
11
19
12
Wave 3
Men, SVORI
Women, SVORI
Juvenile, SVORI
12
Wave 2
26
13
14
11
10
7
Wave 4
Men, SVORI
Women, SVORI
Juvenile, SVORI
Men, non-SVORI
Women, non-SVORI
Juvenile, non-SVORI
Note: Wave 1 = 30 days pre-release; Wave 2 = 3 months post-release; Wave 3 = 9 month post-release; Wave 4 = 15
months post-release.
77
Multi-site Evaluation of SVORI: Summary & Synthesis
Summary
Key points with respect to service needs and receipt include the
following:
78
ƒ
Expressed service needs were high and similar between
the SVORI and non-SVORI groups for most items.
ƒ
Expressed service needs declined between the prerelease interview and the interview conducted 3 months
post-release, remaining relatively stable thereafter.
ƒ
Expressed service needs were highest for the adult
females and lowest for the juvenile males.
ƒ
Expressed service needs were highest for
employment/education/skills services, with need for
education and employment services being reported by
substantial proportions of all groups at each interview.
ƒ
Adult SVORI programs were successful in significantly
increasing the delivery of most services to program
participants, particularly in prison.
ƒ
The four juvenile SVORI programs, on average, did not
provide more services to program participants as
measured by the service bundle scores. SVORI program
participants were more likely, before release, to have
had help finding a place to live (31%, as opposed to
18%) and to have received employment-related services
(42%, as opposed to 27%). SVORI program participants
were also about twice as likely to report receiving
educational and employment-related services in the 3
months after release.
ƒ
Reported receipt of services declined dramatically after
release for all groups.
ƒ
Levels of expressed need were higher—sometimes
substantially higher—than reports of service receipt, for
most service items.
Outcomes
The SVORI programs were intended to provide programs and
services that would result in improvements in community
involvement, employment, family, health (including mental
health), housing, substance use, and criminal behavior,
including supervision compliance. Extensive measurement of
these outcomes was included in the three waves of post-release
interviews, augmented by administrative data describing postrelease arrest and reincarceration. This chapter provides
information on outcomes across key domain areas, focusing on
outcomes in housing, employment, substance use, and criminal
behavior. As in previous chapters, the outcomes are presented
as weighted means generated using Proc Survey Means with
propensity-score-based weights, with tests of significance
generated within the framework of logistic or regression models
of outcomes as a function of the SVORI indicator (estimated
with Proc Survey Logistic or Proc Survey Regression, using SAS
9.1.3).
The adult male respondents are described first, followed by the
adult female respondents and juvenile male respondents. The
final section in this chapter assesses similarities and differences
among the three groups. Housing, then employment, then
substance use, and then criminal behavior outcomes are
discussed.
ADULT MALE OUTCOMES
Three dimensions of housing were examined as reentry
outcomes—housing independence (defined as living in one’s
own house or apartment, contributing to the costs of housing,
or having one’s name on the lease or mortgage of the current
residence), housing stability (defined as having lived in only
79
Multi-site Evaluation of SVORI: Summary & Synthesis
one place during the reference period or two places if the move
was to secure one’s own place or a nicer place), and the extent
of challenges faced in locating housing (respondents were
classified as not having housing challenges if they were not
homeless, reported that they did not have trouble finding a
place to live, and reported that their current living situation was
better or about the same as their last one).The SVORI and nonSVORI groups were compared on these outcomes at 3, 9, and
15 months post-release.
Exhibit 32 shows the weighted proportion of men in each
group, with parameter estimates, probability values, and odds
ratios from the logistic regression models. As can be seen, the
SVORI and non-SVORI groups are similar along these core
housing dimensions, indicating that SVORI programming did
not significantly improve the post-release housing experiences
for adult male prisoners returning to their communities. The
single statistically significant difference (at the 0.05 level) was
that the SVORI group, on average, was more likely to report
being housing-independent than the non-SVORI group
15 months after release.
Exhibit 32. Weighted
means and regression
parameter estimates for
housing outcomes
(adult males)
Wave 2
Housing independence
Housing stability
No housing challenges
Wave 3
Housing independence
Housing stability
No housing challenges
Wave 4
Housing independence
Housing stability
No housing challenges
SVORI
Mean
N = 529
0.723
0.784
0.837
N = 565
0.818
0.695
0.847
N = 582
0.861
0.672
0.815
NonSVORI
Mean
N = 455
0.692
0.781
0.815
N = 470
0.829
0.709
0.820
N = 531
0.798
0.728
0.833
Par.
Est.
p
value
OR
0.151
0.017
0.154
0.290
0.915
0.365
1.163
1.017
1.167
−0.074
−0.070
0.201
0.669
0.621
0.252
0.929
0.932
1.222
0.450*
−0.267
−0.123
0.012
0.070
0.086
1.569
0.766
0.884
Note: Par. Est. = parameter estimate for weighted regression; OR = odds ratio.
N’s are the total responses for each wave of interviews and do not reflect any
item missingness. Wave 2 = 3 months post-release; Wave 3 = 9 month postrelease; Wave 4 = 15 months post-release.
*p < 0.05
80
Outcomes
Employment was a primary focus of many of the SVORI
programs (confirmed by the higher employment-related
service-receipt scores consistently reported by the SVORI
group). Employment was also of considerable importance to the
respondents, who consistently indicated high levels of need for
services to improve their employment, education, and other
skills. Extensive data were collected from respondents to assess
their post-release employment experiences.
Exhibit 33 shows the results for key employment outcomes.
These results suggest that SVORI program participants had
better post-release employment experiences than the nonSVORI comparison respondents—if only moderately so. They
were more likely to report that they were currently supporting
themselves with a job at 3 and 15 months post-release. They
reported working about the same number of months on
average—about two thirds of the available months (2 of 3
months immediately after release, and about 4 of 6 months at
the 9- and 15-month interviews)—and were equally likely to
have reported working all months in the reference period.
SVORI participants appear to have secured better jobs—jobs
that provided formal pay and benefits.
Exhibit 33. Weighted means and regression parameter estimates for employment outcomes
(adult males)
Wave 2
Supported self with job
Number months worked
Worked each month
Received formal pay
Job benefits
Wave 3
Supported self with job
Number months worked
Worked each month
Received formal pay
Job benefits
Wave 4
Supported self with job
Number months worked
Worked each month
Received formal pay
Job benefits
SVORI Mean
N = 529
0.64
2.04
0.38
0.84
0.47
N = 565
0.68
3.83
0.43
0.80
0.53
N = 582
0.71
3.70
0.44
0.78
0.52
Non-SVORI Mean
N = 455
0.59
1.96
0.39
0.74
0.39
N = 470
0.68
3.73
0.44
0.77
0.42
N = 531
0.60
3.50
0.42
0.74
0.44
Par. Est.
p Value
OR
0.230
0.079
−0.056
0.604
0.337
0.086
0.420
0.716
0.001
0.028
1.258
NA
0.945
1.829
1.400
0.027
0.102
−0.033
0.178
0.472
0.848
0.536
0.823
0.310
0.001
1.027
NA
0.968
1.195
1.602
0.481
0.197
0.045
0.183
0.326
0.001
0.252
0.772
0.306
0.034
1.617
NA
1.046
1.201
1.386
Note: Par. Est. = parameter estimate for weighted regression; OR = odds ratio; NA = not applicable, because model
was a regression. N’s are the total responses for each wave of interviews and do not reflect any item missingness.
Wave 2 = 3 months post-release; Wave 3 = 9 month post-release; Wave 4 = 15 months post-release.
81
Multi-site Evaluation of SVORI: Summary & Synthesis
The levels of drug use
were quite high. At 15
months post-release,
less than 50% of both
groups had been
abstinent in the in the
past 30 days, according
to self-reported use and
a negative drug test.
Substance use outcomes were measured both by self-report
during all follow-up interviews and by oral fluids drug tests
administered to nonincarcerated respondents at the 3- and 15month interviews. The results for the core substance use
outcomes are shown in Exhibit 34. Self-reported substance use
was generally lower for the SVORI group than for the nonSVORI group, and in several cases these differences were
statistically significant. Similarly, the outcome that reflects
either self-reported drug use over the past 30 days or
confirmed (by drug tests) drug use, the results again suggest
that the SVORI participants were less likely to use drugs,
although these differences are not statistically significant.
Overall, the results suggest that SVORI program participants
were doing somewhat better with respect to drug use but that
all men continued to be at high risk for continued drug use.
The SVORI logic model suggests that services responsive to
needs will result in improvements in intermediate outcomes,
including housing, employment, and substance use.
Improvements in these outcomes, in turn, are hypothesized to
result in improvements in criminal behavior. Because of the
importance of recidivism, multiple measures were included in
the evaluation to determine program effects on desistance from
criminal activity. These measures include self-reported and
official measures of criminal behavior. Core criminalbehavior/recidivism outcomes based on self-reports are shown
in Exhibit 35.
The first measure in the exhibit does not directly measure
criminal behavior, but rather perpetration of violence.
Respondents were asked about several specific types of
violence: threatening to hit, throwing, pushing/grabbing/
shoving, slapping/kicking/biting/hitting, and threatening to use
or using a weapon. The measure was scored 1 if the respondent
answered “yes” to any of these queries and was zero,
otherwise. The men in the SVORI group were slightly less likely
to report having perpetrated violence than the men in the
comparison group at each post-release time period, but none of
the differences was statistically significant.
82
Outcomes
Exhibit 34. Weighted means and regression parameter estimates for substance use
outcomes (adult males)
SVORI
Mean
NonSVORI
Mean
N = 529
N = 455
No self-reported drug use
0.74
No self-reported drug use other than marijuana
or steroids
Par.
Est.
p Value
OR
0.70
0.170
0.243
1.185
0.85
0.85
−0.030
0.868
0.970
No self-reported drug use past 30 days
0.79
0.77
0.160
0.313
1.174
No self-reported drug use other than marijuana
or steroids past 30 days
0.88
0.89
−0.066
0.747
0.936
No self-reported drug use past 30 days or
positive drug test
0.54
0.52
0.093
0.475
1.098
N = 565
N = 470
No self-reported drug use
0.57
0.52
0.201
0.125
1.223
No self-reported drug use other than marijuana
or steroids
0.74
0.71
0.177
0.227
1.194
No self-reported drug use past 30 days
0.69
0.62
0.301
0.028
1.351
No self-reported drug use other than marijuana
or steroids past 30 days
0.81
0.79
0.088
0.589
1.092
—
—
—
—
—
N = 582
N = 531
No self-reported drug use
0.58
0.50
0.337
0.012
1.401
No self-reported drug use other than marijuana
or steroids
0.78
0.72
0.311
0.045
1.365
No self-reported drug use past 30 days
0.66
0.62
0.1512
0.2784
1.163
No self-reported drug use other than marijuana
or steroids past 30 days
0.82
0.80
0.168
0.326
1.182
No self-reported drug use past 30 days or
positive drug test
0.46
0.43
0.118
0.381
1.126
Wave 2
Wave 3
No self-reported drug use past 30 days or
positive drug test
Wave 4
Note: Par. Est. = parameter estimate for weighted regression; OR = odds ratio. N’s are the total responses for each
wave of interviews and do not reflect any item missingness. Wave 2 = 3 months post-release; Wave 3 = 9 month
post-release; Wave 4 = 15 months post-release.
83
Multi-site Evaluation of SVORI: Summary & Synthesis
Exhibit 35. Weighted means and regression parameter estimates for core self-reported
recidivism outcomes (adult males)
SVORI
Mean
NonSVORI
Mean
N = 529
N = 455
No perpetration of violence
0.73
Complied with conditions of supervision
Par. Est.
p Value
OR
0.71
0.117
0.422
1.124
0.78
0.78
0.005
0.977
1.005
No criminal behavior
0.79
0.73
0.327
0.034
1.386
No violent or weapons crimes
0.90
0.91
−0.076
0.732
0.927
Not reincarcerated at follow-up
0.93
0.92
0.235
0.352
1.265
N = 565
N = 470
No perpetration of violence
0.64
0.60
0.178
0.175
1.195
Complied with conditions of supervision
0.69
0.70
−0.025
0.881
0.975
No criminal behavior
0.64
0.59
0.207
0.112
1.230
No violent or weapons crimes
0.85
0.82
0.222
0.199
1.249
Not reincarcerated at follow-up
0.73
0.74
−0.067
0.641
0.935
N = 582
N = 531
No perpetration of violence
0.69
0.67
0.069
0.595
1.072
Complied with conditions of supervision
0.66
0.57
0.398
0.023
1.489
No criminal behavior
0.66
0.61
0.189
0.136
1.208
No violent or weapons crimes
0.84
0.83
0.073
0.654
1.076
Not reincarcerated at follow-up
0.64
0.66
−0.065
0.612
0.937
Wave 2
Wave 3
Wave 4
Note: Par. Est. = parameter estimate for weighted regression. OR = odds ratio. N’s are the total responses for each
wave of interviews and do not reflect any item missingness. Wave 2 = 3 months post-release; Wave 3 = 9 month
post-release; Wave 4 = 15 months post-release.
The second core measure of criminal behavior/recidivism was
compliance with conditions of supervision, important because
the majority of men reported being under post-release
supervision throughout the follow-up period. As shown in
Exhibit 35, the results were mixed: There was no difference in
reports of compliance at 3 months; at 9 months, slightly fewer
members of the SVORI group reported complying; and, at 15
months, a significantly higher percentage of the SVORI group
reported complying with the conditions of their supervision.
Exhibit 35 also includes self-reports of any criminal behavior
(which includes violent crimes, carrying a weapon, other crimes
against people, drug possession crimes, drug sales crimes,
driving while intoxicated, property crimes, and lesser types of
crimes, such as prostitution, soliciting, shoplifting, or disorderly
84
Outcomes
Three months after
release, about 20%
were either
reincarcerated or had
been booked into jail or
prison for 24 hours or
more . . . a percentage
that increased to 50%
at 15 months.
conduct) and self-reported involvement in violent or weapons
offenses. SVORI program participants were more likely than
non-SVORI comparison respondents to report committing no
crimes since release, or since the previous interview. This
difference is statistically significant at the 3-month interview
(p < 0.05), but not for subsequent follow-up periods. About 9–
10% of both groups reported committing a violent crime,
carrying a weapon, or both in the 3 months after their release
from prison. In subsequent periods, greater numbers reported
either committing a violent crime or carrying a weapon since
the previous interview. None of the differences was statistically
significant (the p value for the Wave 3 SVORI to non-SVORI
difference was 0.2). The final core criminal-behavior/recidivism
outcome based on self-reported data was whether the
respondent was not reincarcerated at the time of his follow-up
interview. The two groups were similar, with nearly 10%
reincarcerated within 3 months of release. By the time of the 9month interview, more than a quarter were in prison, and more
than one third were reincarcerated at the 15-month interview.
The remaining set of criminal recidivism measures were based
on official data sources and therefore reflect criminal behavior
detected by authorities. These measures include both rearrest
(obtained from NCIC and processed as described in Lattimore
and Steffey, 2009) and reincarceration in state prisons (obtained
from the state Departments of Corrections). In contrast to the
self-report measures, these data were successfully obtained for
almost all respondents. As with the self-report measures, the
reported means are weighted, with the use of the propensity
scores, and the parameter estimates are for the SVORI indicator
variable in the weighted logistic regression models. These core
recidivism measures based on official records are shown in
Exhibit 36.
85
Multi-site Evaluation of SVORI: Summary & Synthesis
Exhibit 34. Official measures of recidivism (adult males)
Rearrest
First rearrest within 3 months of release
First rearrest within 6 months of release
First rearrest within 9 months of release
First rearrest within 12 months of release
First rearrest within 15 months of release
First rearrest within 21 months of release
First rearrest within 24 months of release
Rearrest within 21 months for violent crime
Rearrest within 21 months for property crime
Rearrest within 21 months for drug crime
Rearrest within 21 months for public order crime
Rearrest within 21 months for other crime
Rearrest within 24 months for violent crime
Rearrest within 24 months for property crime
Rearrest within 24 months for drug crime
Rearrest within 24 months for public order crime
Rearrest within 24 months for other crime
Reincarceration
First reincarceration within 3 months of release
First reincarceration within 6 months of release
First reincarceration within 9 months of release
First reincarceration within 12 months of release
First reincarceration within 15 months of release
First reincarceration within 21 months of release
First reincarceration within 24 months of release
SVORI
Mean
N = 806
16%
28%
41%
49%
55%
64%
68%
19%
23%
28%
41%
3%
20%
26%
30%
44%
3%
N = 863
3%
11%
19%
25%
30%
39%
42%
NonSVORI
Mean
N = 775
18%
32%
44%
51%
56%
66%
71%
21%
24%
30%
45%
6%
23%
27%
32%
49%
6%
N = 834
4%
10%
17%
25%
29%
36%
39%
Par. Est.
SE
OR
−0.163
−0.191
−0.129
−0.089
−0.036
−0.089
−0.131
−0.112
−0.080
−0.118
−0.175
−0.560
−0.142
−0.054
−0.117
−0.189
−0.585
0.136
0.111
0.104
0.102
0.103
0.107
0.112
0.129
0.120
0.114
0.104
0.250
0.127
0.117
0.112
0.104
0.249
0.849
0.826
0.879
0.915
0.964
0.914
0.877
0.894
0.923
0.889
0.839
0.571
0.867
0.948
0.890
0.828
0.557
−0.163
0.062
0.113
−0.023
0.033
0.135
0.128
0.280
0.160
0.128
0.114
0.108
0.102
0.102
0.849
1.064
1.120
0.977
1.033
1.145
1.137
Note: Par. Est. = parameter estimate for weighted regression; OR = odds ratio; SE = standard error. For rearrest,
SVORI N = 806 for all periods, except N = 787 for 24-month measures; non-SVORI N = 775 for all periods, except
N = 759 for 24-month measures. For reincarceration, SVORI N = 863 for all periods; non-SVORI N = 834, except
for N = 817 for 24-month measure.
*p < 0.05 for test of significant difference between SVORI and non-SVORI.
The findings suggest that (1) members of the SVORI group
were less likely to be rearrested across the 24 months after
release than the non-SVORI group, although the differences
were not statistically significant; and (2) rearrest rates for
these serious and violent offenders were quite high, with about
70% having had at least one new arrest within 24 months of
release.
The findings for reincarceration indicate that the SVORI and
non-SVORI groups were equally likely to be reincarcerated
86
Outcomes
throughout the 24-month follow-up period—and rates were
actually higher for SVORI participants (although not
significantly so) after 3 months. These results are somewhat at
odds with both the self-reported data and the arrest data,
which consistently, if weakly, suggest less criminal activity by
the SVORI participants.
ADULT FEMALE OUTCOMES
Exhibit 37 shows the weighted proportion of women in each
group for the core housing outcomes, with estimates,
probability values, and odds ratios from the logistic regression
models. As can be seen, the SVORI and non-SVORI groups are
similar, indicating that SVORI programming did not significantly
improve the post-release housing experiences for adult female
prisoners returning to their communities. The temporal patterns
suggest that housing independence improved gradually over
the post-release follow-up period, but housing stability declined
over time. The pattern observed for the measure of housing
challenges, which is perhaps the broadest measure of difficulty
in finding quality housing, indicates that the time period from 3
to 9 months post-release was the period during which women
experienced the most challenges, with their situation appearing
to improve by the 15-month post-release time period.
Exhibit 37. Weighted
means and regression
parameter estimates for
housing outcomes
(adult females)
Wave 2
Housing independence
Housing stability
No housing challenges
Wave 3
Housing independence
Housing stability
No housing challenges
Wave 4
Housing independence
Housing stability
No housing challenges
SVORI
Mean
N = 110
0.65
0.76
0.82
N = 119
0.78
0.64
0.72
N = 124
0.82
0.65
0.82
NonSVORI
Mean
N = 134
0.69
0.79
0.77
N = 134
0.77
0.73
0.74
N = 152
0.73
0.65
0.76
Par.
Est.
p
value
OR
−0.18
−0.17
0.26
0.52
0.60
0.43
0.83
0.85
1.30
0.05
−0.41
−0.08
0.88
0.16
0.80
1.05
0.67
0.93
0.52
0.01
0.38
0.13
0.98
0.26
1.68
1.01
1.46
Note: Par. Est. = parameter estimate for weighted regression. OR = odds ratio;
N’s are the total responses for each wave of interviews and do not reflect any
item missingness. Wave 2 = 3 months post-release; Wave 3 = 9 month postrelease; Wave 4 = 15 months post-release.
* p < 0.05
87
Multi-site Evaluation of SVORI: Summary & Synthesis
Exhibit 38 shows the results for key employment outcomes.
Several findings suggest more positive employment outcomes
for women who enrolled in SVORI programs. Notably, the
women who participated in SVORI programming were
significantly more likely than the non-SVORI group to report
having supported themselves with a job at the 15-month time
period. Findings also indicate that, at the 15-month time
period, women who had enrolled in SVORI programs had
worked significantly more months since the last interview than
comparison group members. In addition, the SVORI group was
significantly more likely to report receiving formal pay for their
job and slightly more likely to report that their job provided
benefits at the 3-month post-release time period (p = 0.07).
Exhibit 38. Weighted
means and regression
parameter estimates for
employment outcomes
(adult females)
Wave 2
Supported self with job
Number months worked
Worked each month
Received formal pay
Job benefits
Wave 3
Supported self with job
Number months worked
Worked each month
Received formal pay
Job benefits
Wave 4
Supported self with job
Number months worked
Worked each month
Received formal pay
Job benefits
SVORI
Mean
N = 110
0.53
1.57
0.24
0.89
0.41
N = 119
0.61
3.72
0.42
0.90
0.41
N = 124
0.68
3.74
0.44
0.90
0.42
NonSVORI
Par.
Mean
Est.
N = 134
0.49
0.16
1.38
0.19
0.32
−0.40
0.74
1.04
0.26
0.68
N = 134
0.56
0.17
3.25
0.47
0.46
−0.17
0.73
1.16
0.33
0.34
N = 152
0.45
0.95
2.93
0.81
0.42
0.12
0.74
1.09
0.37
0.23
p
Value
OR
0.56
0.31
0.30
0.03
0.07
1.17
NA
0.67
2.83
1.98
0.54
0.19
0.60
0.00
0.29
1.18
NA
0.85
3.20
1.41
0.00
0.02
0.71
0.02
0.48
2.59
NA
1.13
2.97
1.26
Note: Par. Est. = parameter estimate for weighted regression; OR = odds ratio;
NA = not applicable, because model was a regression. N’s are the total
responses for each wave of interviews and do not reflect any item
missingness. Wave 2 = 3 months post-release; Wave 3 = 9 month postrelease; Wave 4 = 15 months post-release.
88
Outcomes
The patterns for
substance use, based
on the combined selfreport and drug test
measures, suggest
increasing substance
use over time for both
groups.
The results for the core substance use outcomes are shown in
Exhibit 39. Self-reported substance use was generally lower for
the SVORI group than for the non-SVORI group; however,
these differences were not statistically significant (although
past-30-day drug use at the 15-month interview was marginally
significant, p < 0.10). The outcome that reflects either selfreported or confirmed (by drug tests) use shows that the
SVORI participants were significantly less likely to use drugs
from release to 3 months post-release (p < 0.01), and, when
the measure was limited to past-30-day use, less likely to have
used during the previous 30 days at both the 3- and 15-month
post-release time periods. With respect to use of specific drugs,
the only consistent difference between the SVORI and nonSVORI groups was that, at 15 months post-release, the SVORI
group self-reported significantly lower cocaine use than the
non-SVORI group (87%, as opposed to 73%; p < 0.05). The
drug test results also confirmed lower cocaine use among the
SVORI participants, at both 3 and 15 months post-release (data
not shown).
Findings for the core self-reported criminal-behavior/recidivism
outcomes are shown in Exhibit 40. The women in the SVORI
group were less likely to report violence perpetration than the
women in the comparison group at the 15-month post-release
time period; however, among the women who were under postrelease supervision, the SVORI group was less likely to report
that they had complied with the conditions of their supervision
during the first 9 months after release.21 The pattern for
supervision compliance reversed at the 15-month time period,
when the SVORI group reported higher compliance than the
non-SVORI group during the previous 6 months.
21
Most women reported being under post-release supervision
throughout the follow-up period: More than three quarters of the
women were under post-release supervision at the time of the 3month post-release interview. At the 9-month interview, 80% of
the SVORI group (compared with 56% of the non-SVORI group)
were still under post-release supervision. At the final interview
wave, 54% of SVORI participants and 43% of comparison women
were currently under post-release supervision.
89
Multi-site Evaluation of SVORI: Summary & Synthesis
Exhibit 39. Weighted means and regression parameter estimates for substance use
outcomes (adult females)
SVORI
Mean
NonSVORI
Mean
N = 110
N = 134
No self-reported drug use
0.78
0.75
No self-reported drug use other than marijuana
or steroids
0.87
No self-reported drug use past 30 days
Par.
Est.
p Value
OR
0.18
0.59
1.19
0.85
0.17
0.67
1.18
0.85
0.81
0.29
0.45
1.34
No self-reported drug use other than marijuana
or steroids past 30 days
0.92
0.87
0.52
0.29
1.69
No self-reported drug use past 30 days or
positive drug test
0.67
0.49
0.76
0.00
2.15
N = 119
N = 134
No self-reported drug use
0.60
0.61
−0.03
0.91
0.97
No self-reported drug use other than marijuana
or steroids
0.69
0.71
−0.11
0.71
0.89
No self-reported drug use past 30 days
0.69
0.73
−0.18
0.56
0.84
No self-reported drug use other than marijuana
or steroids past 30 days
0.76
0.82
−0.37
0.29
0.69
—
—
—
—
N = 124
N = 152
No self-reported drug use
0.63
0.55
0.32
0.24
1.38
No self-reported drug use other than marijuana
or steroids
0.77
0.67
0.53
0.09
1.69
No self-reported drug use past 30 days
0.75
0.63
0.55
0.07
1.74
No self-reported drug use other than marijuana
or steroids past 30 days
0.82
0.74
0.49
0.15
1.63
No self-reported drug use past 30 days or
positive drug test
0.46
0.38
0.33
0.23
1.39
Wave 2
Wave 3
No self-reported drug use past 30 days or
positive drug test
Wave 4
—
Note: Par. Est. = parameter estimate for weighted regression; OR = odds ratio. N’s are the total responses for each
wave of interviews and do not reflect any item missingness. Wave 2 = 3 months post-release; Wave 3 = 9 month
post-release; Wave 4 = 15 months post-release.
90
Outcomes
Exhibit 40. Weighted means and regression parameter estimates for core self-report
recidivism outcomes (adult females)
SVORI
Mean
NonSVORI
Mean
N = 110
N = 134
No perpetration of violence
0.76
Complied with conditions of supervision
Par. Est.
p Value
OR
0.71
0.26
0.41
1.30
0.78
0.83
−0.35
0.35
0.70
No criminal behavior
0.78
0.82
−0.27
0.43
0.76
No violent or weapons crimes
0.93
0.91
0.18
0.74
1.20
Not reincarcerated at follow-up
0.93
0.94
−0.12
0.82
0.89
N = 119
N = 134
No perpetration of violence
0.62
0.60
0.07
0.79
1.08
Complied with conditions of supervision
0.59
0.76
−0.79
0.03
0.45
No criminal behavior
0.69
0.75
−0.31
0.30
0.73
No violent or weapons crimes
0.90
0.93
−0.27
0.56
0.77
Not reincarcerated at follow-up
0.78
0.86
−0.53
0.12
0.59
N = 124
N = 152
No perpetration of violence
0.68
0.56
0.53
0.05
1.70
Complied with conditions of supervision
0.68
0.60
0.35
0.37
1.43
No criminal behavior
0.70
0.64
0.31
0.26
1.36
No violent or weapons crimes
0.96
0.87
1.26
0.01
3.53
Not reincarcerated at follow-up
0.77
0.78
−0.10
0.74
0.91
Wave 2
Wave 3
Wave 4
Note: Par. Est. = parameter estimate for weighted regression; OR = odds ratio. N’s are the total responses for each
wave of interviews and do not reflect any item missingness. Wave 2 = 3 months post-release; Wave 3 = 9 month
post-release; Wave 4 = 15 months post-release.
Exhibit 40 also shows the proportions of each group that
reported having committed no criminal behavior and no violent
or weapons crimes in the period since release or last interview.
Differences in any crime are not statistically significant, but
SVORI program participants were less likely to report no
criminal behavior at Waves 2 and 3 and more likely to report no
criminal behavior at Wave 4. SVORI program participants were
also more likely to report no violent/weapons crimes at Waves
2 and 4.
The final criminal-behavior/recidivism outcome considered here
is an indicator of whether the respondent was reincarcerated at
the time of the follow-up interview. No significant differences on
these outcomes were observed between SVORI and comparison
women for any follow-up period.
91
Multi-site Evaluation of SVORI: Summary & Synthesis
The criminal recidivism measures based on official records are
shown in Exhibit 41. The findings for rearrest indicate that the
women in the SVORI and non-SVORI groups were equally likely
to be rearrested within 3 and 6 months of release but that the
SVORI participants were significantly less likely to be rearrested
within 9, 12, 15, and 21 months of release. When type of
rearrest (person/violent crimes, property crimes, drug crimes,
public order crimes, and other crimes) was examined at the 21and 24-month time periods, no significant differences were
evident.
Exhibit 41. Official measures of recidivism (adult females)
Rearrest
First rearrest within 3 months of release
First rearrest within 6 months of release
First rearrest within 9 months of release
First rearrest within 12 months of release
First rearrest within 15 months of release
First rearrest within 21 months of release
First rearrest within 24 months of release
Rearrest within 21 months for violent crime
Rearrest within 21 months for property crime
Rearrest within 21 months for drug crime
Rearrest within 21 months for public-order crime
Rearrest within 21 months for other crime
Rearrest within 24 months for violent crime
Rearrest within 24 months for property crime
Rearrest within 24 months for drug crime
Rearrest within 24 months for public-order crime
Rearrest within 24 months for other crime
Reincarceration
First reincarceration within 3 months of release
First reincarceration within 6 months of release
First reincarceration within 9 months of release
First reincarceration within 12 months of release
First reincarceration within 15 months of release
First reincarceration within 21 months of release
First reincarceration within 24 months of release
SVORI
Mean
N = 143
12%
18%
25%
29%
33%
44%
49%
7%
17%
15%
32%
4%
8%
19%
16%
34%
2%
N = 153
4%
10%
15%
24%
30%
36%
41%
NonSVORI
Mean
N = 194
14%
26%
36%
42%
49%
59%
60%
12%
22%
21%
40%
7%
12%
24%
21%
41%
7%
N = 204
2%
8%
11%
14%
17%
21%
22%
Par. Est.
p Value
OR
−0.16
−0.46
−0.52
−0.56
−0.67
−0.63
−0.46
−0.62
−0.26
−0.41
−0.36
−0.59
−0.48
−0.31
−0.36
−0.30
−1.11
0.64
0.12
0.04
0.02
0.01
0.01
0.05
0.18
0.37
0.18
0.14
0.31
0.23
0.27
0.23
0.21
0.06
0.85
0.63
0.59
0.57
0.51
0.53
0.63
0.54
0.77
0.66
0.70
0.56
0.62
0.73
0.69
0.74
0.33
0.76
0.24
0.38
0.69
0.75
0.74
0.87
0.32
0.54
0.25
0.02
0.00
0.00
0.00
2.13
1.27
1.46
1.99
2.13
2.09
2.38
Note: Par. Est. = parameter estimate for weighted regression; OR = odds ratio. For rearrest, SVORI N = 143 for all
periods, except N = 140 for 24-month measures; non-SVORI N = 194 for all periods, except N = 193 for 24 month
measures. For reincarceration, SVORI N = 153 for all periods, except N = 150 for 24-month measure; non-SVORI
N = 204, except N = 203 for 24-month measure.
92
Outcomes
The findings for reincarceration indicate that the SVORI and
non-SVORI groups were equally likely to be reincarcerated
within 3, 6, and 9 months of release but that the SVORI
participants were significantly more likely to be reincarcerated
within 12, 15, 21, and 24 months of release. Considered in
conjunction with the arrest findings—of significantly less
likelihood of arrest for SVORI program participants—these
results are somewhat puzzling. Indeed, 24 months after release
49% of SVORI program participants had an arrest, and 41%
had been reincarcerated. In comparison, 60% of non-SVORI
respondents had been arrested, but only 22% had been
reincarcerated. Preliminary investigations failed to reveal any
obvious explanation for the findings (e.g., percentage on
supervision who were at risk for revocation).
JUVENILE MALE OUTCOMES
Exhibit 42 shows the core housing outcomes (housing
independence, housing stability, and the extent of challenge in
locating housing after release) for the juvenile males. As can be
seen, the SVORI and non-SVORI groups are similar, indicating
that SVORI programming did not significantly improve the postrelease housing experiences for returning adult female
prisoners. The temporal patterns suggest that housing
independence improved gradually over the post-release followup period. At 15 months post-release, non-SVORI juvenile
males had significantly more housing independence than SVORI
juvenile males.
93
Multi-site Evaluation of SVORI: Summary & Synthesis
Exhibit 42. Weighted
means and regression
parameter estimates for
housing outcomes
(juvenile males)
Wave 2
Housing independence
Housing stability
No housing challenges
Wave 3
Housing independence
Housing stability
No housing challenges
Wave 4
Housing independence
Housing stability
No housing challenges
SVORI
Mean
N = 105
0.23
0.81
0.88
N = 108
0.35
0.62
0.94
N = 107
0.37
0.74
0.93
NonSVORI
Mean
N = 131
0.31
0.78
0.92
N = 131
0.47
0.70
0.88
N = 141
0.51
0.71
0.90
Par.
Est.
p
value
OR
−0.43
0.16
−0.47
0.16
0.65
0.30
0.65
0.21
0.63
−0.50
−0.37
0.70
0.09
0.25
0.21
0.61
0.69
2.01
−0.61
0.17
0.31
0.04
0.61
0.59
0.54
1.18
1.37
Note: Par. Est. = parameter estimate for weighted regression; OR = odds ratio.
N’s are the total responses for each wave of interviews and do not reflect any
item missingness. Wave 2 = 3 months post-release; Wave 3 = 9 month postrelease; Wave 4 = 15 months post-release.
* p < 0.05
Exhibit 43 shows the results for key employment outcomes.
There are few statistically significant differences between the
two groups. More non-SVORI respondents reported supporting
themselves with a job at the first two post-release interviews,
but the differences were not statistically significant. The
proportions of both groups reporting that they had supported
themselves with a job increased for both groups at the 15month interview, but increased more for the SVORI group so
that more SVORI program participants reported supporting
themselves with a job. The number of months worked
increased over time for both groups. Employed SVORI group
members were more likely to report that they received formal
pay at all waves and were more likely to report that their job
offered benefits at Waves 3 and 4.
94
Outcomes
Exhibit 43. Weighted
means and regression
parameter estimates for
employment outcomes
(juvenile males)
Wave 2
Supported self with job
Number months worked
Worked each month
Received formal pay
Job benefits
Wave 3
Supported self with job
Number months worked
Worked each month
Received formal pay
Job benefits
Wave 4
Supported self with job
Number months worked
Worked each month
Received formal pay
Job benefits
SVORI
Mean
N = 105
0.32
1.04
0.18
0.70
0.25
N = 108
0.32
2.20
0.22
0.83
0.45
N = 107
0.53
2.94
0.28
0.76
0.53
NonSVORI
Mean
N = 131
0.40
1.21
0.22
0.65
0.37
N = 131
0.39
2.56
0.24
0.66
0.42
N = 141
0.44
2.49
0.27
0.71
0.40
Par.
Est.
p
Value
OR
−0.35
−0.18
−0.27
0.25
−0.58
0.25
0.35
0.60
0.56
0.18
0.70
NA
0.77
1.28
0.56
−0.31
−0.36
−0.10
0.93
0.13
0.30
0.30
0.80
0.04
0.74
0.74
NA
0.90
2.53
1.14
0.37
0.45
0.02
0.29
0.80
0.22
0.22
0.96
0.47
0.02
1.45
NA
1. 01
1.34
2.22
Note: Par. Est. = parameter estimate for weighted regression; OR = odds ratio;
NA = not applicable, because model was a regression. N’s are the total
responses for each wave of interviews and do not reflect any item
missingness. Wave 2 = 3 months post-release; Wave 3 = 9 month postrelease; Wave 4 = 15 months post-release.
The results for the core substance use outcomes are shown in
Exhibit 44. Self-reported substance use was generally lower for
the SVORI group than for the non-SVORI group immediately
after release (i.e., at Wave 2); however, the reverse was true
by 15 months after release (these differences were not
statistically significant). SVORI program participants were
significantly more likely 3 months post-release to have both no
self-reported drug use in the past 30 days and a negative urine
test (62%, as opposed to 48%). Over time, the proportions of
both groups who reported drug use and who tested positive
increased.
95
Multi-site Evaluation of SVORI: Summary & Synthesis
Exhibit 44. Weighted means and regression parameter estimates for substance use
outcomes (juvenile males)
SVORI
Mean
NonSVORI
Mean
N = 110
N = 134
No self-reported drug use
0.79
No self-reported drug use other than marijuana
or steroids
Par.
Est.
p Value
0.73
0.33
0.32
1.4
0.95
0.91
0.62
0.28
1.86
No self-reported drug use past 30 days
0.85
0.79
0.38
0.3
1.46
No self-reported drug use other than marijuana
or steroids past 30 days
0.96
0.92
0.71
0.27
2.03
No self-reported drug use past 30 days or
positive drug test
0.62
0.48
0.6
0.04
1.83
N = 119
N = 134
No self-reported drug use
0.59
0.56
0.12
0.69
1.13
No self-reported drug use other than marijuana
or steroids
0.87
0.88
−0.09
0.83
0.91
No self-reported drug use past 30 days
0.66
0.66
0.01
1
1
No self-reported drug use other than marijuana
or steroids past 30 days
0.93
0.94
−0.15
0.79
0.86
No self-reported drug use past 30 days or
positive drug test
—
—
—
—
N = 124
N = 152
No self-reported drug use
0.56
0.58
−0.08
0.8
0.93
No self-reported drug use other than marijuana
or steroids
0.84
0.9
−0.52
0.22
0.59
No self-reported drug use past 30 days
0.61
0.71
−0.46
0.13
0.63
No self-reported drug use other than marijuana
or steroids past 30 days
0.9
0.93
−0.32
0.53
0.73
No self-reported drug use past 30 days or
positive drug test
0.46
0.43
0.12
0.7
1.13
Wave 2
Wave 3
Wave 4
OR
—
Note: Par. Est. = parameter estimate for weighted regression; OR = odds ratio. N’s are the total responses for each
wave of interviews and do not reflect any item missingness. Wave 2 = 3 months post-release; Wave 3 = 9 month
post-release; Wave 4 = 15 months post-release.
96
Outcomes
Findings for the core self-report criminal-behavior/recidivism
outcomes are shown in Exhibit 45. There were no differences
between the groups on any of these measures. Self-reports of
not perpetrating violence suggested that both groups were
likely to engage in substantial violent acts, with only between
47% and 62% of a group reporting not engaging in violence
over the follow-up period. There was also no difference
between groups in compliance with conditions of supervision—
but reports of compliance declined over time for both groups.
Exhibit 45. Weighted means and regression parameter estimates for core self-report
recidivism outcomes (juvenile males)
SVORI
Mean
NonSVORI
Mean
N = 110
N = 134
No perpetration of violence
0.57
0.62
Complied with conditions of supervision
0.80
No criminal behavior
p Value
OR
−0.21
0.47
0.81
0.86
−0.39
0.32
0.68
0.53
0.48
0.21
0.80
1.23
No violent or weapons crimes
0.76
0.75
0.08
0.96
1.08
Not reincarcerated at follow-up
0.92
0.92
0.01
0.98
1.01
N = 119
N = 134
No perpetration of violence
0.49
0.47
0.10
0.74
1.10
Complied with conditions of supervision
0.71
0.74
−0.17
0.72
0.85
No criminal behavior
0.41
0.41
−0.02
0.54
0.98
No violent or weapons crimes
0.54
0.58
−0.18
0.41
0.83
Not reincarcerated at follow-up
0.73
0.82
−0.50
0.16
0.61
N = 124
N = 152
No perpetration of violence
0.57
0.55
0.09
0.75
1.09
Complied with conditions of supervision
0.66
0.77
−0.56
0.39
0.57
No criminal behavior
0.37
0.38
−0.05
0.50
0.95
No violent or weapons crimes
0.62
0.57
0.19
0.57
1.21
Not reincarcerated at follow-up
0.73
0.80
−0.37
0.26
0.69
Wave 2
Wave 3
Wave 4
Par. Est.
Note: Par. Est. = parameter estimate for weighted regression; OR = odds ratio. N’s are the total responses for each
wave of interviews and do not reflect any item missingness. Wave 2 = 3 months post-release; Wave 3 = 9 month
post-release; Wave 4 = 15 months post-release.
Exhibit 45 also shows the proportions of each group that
reported having committed no criminal behavior and no violent
or weapons crimes in the period since release or last interview.
Again, reports from both groups are similar. What is most
striking about these findings is the level of self-reported
97
Multi-site Evaluation of SVORI: Summary & Synthesis
criminal activity. At 15 months after release, only about 37% of
both groups reported that they had not engaged in criminal
activities since the last interview—suggesting that almost two
thirds (63%) had engaged in criminal activities. With respect to
violent/weapons crimes, both groups were less likely to report
not having committed these types of crimes during the
reference period at the Wave 3 (9 months post-release)
interview. Members of the SVORI group were more likely to
have been incarcerated at the time of the Waves 3 and 4
interviews, although differences were not statistically
significant.
Official records criminal recidivism measures are still being
cleaned and are not yet available.
COMPARISON OF ADULT MALE, ADULT
FEMALE AND JUVENILE MALE OUTCOMES
The results suggest modest improvements in a variety of
outcomes for the adult SVORI participants and few differences
between the juvenile SVORI and non-SVORI participants. Here
some of the key outcomes in housing, employment, substance
use, and criminal behavior are compared for the three groups.
Housing independence
improved over time for all
groups.
98
Housing independence was defined as living in one’s own house
or apartment, contributing to the costs of housing, or having
one’s name on the lease or mortgage of one’s current
residence. Exhibit 46 shows the percentage rated housingindependent for each group at each wave. As can be seen,
housing independence improved over time for all groups. Adult
males and females were similar and much more likely than the
juvenile males to report housing independence. Men who
participated in SVORI programs were significantly more likely
15 months after release to report housing independence (p =
0.01). Juvenile males who were not in a SVORI program were
more likely throughout to report housing independence, and the
difference was significant at 15 months post-release (p = 0.04).
More women who had participated in SVORI programs reported
housing independence at 15 months (82%, as opposed to
73%), but the difference was not statistically significant (p =
0.13).
Outcomes
Exhibit 46. Housing independence by demographic group, study group, and data collection
wave
100%
80%
72%
69%
69%
65%
60%
40%
23%
31%
20%
0%
Wave 2
100%
Men, SVORI
Men, non-SVORI
Women, SVORI
Women, non-SVORI
Juvenile, SVORI
Juvenile, non-SVORI
82%
83%
80%
78%
77%
60%
47%
35%
40%
20%
0%
Wave 3
100%
Men, SVORI
Men, non-SVORI
Women, SVORI
Women, non-SVORI
Juvenile, SVORI
Juvenile, non-SVORI
86%
80%
82%
73%
80%
51%
60%
37%
40%
20%
0%
Wave 4
Men, SVORI
Men, non-SVORI
Women, SVORI
Women, non-SVORI
Juvenile, SVORI
Juvenile, non-SVORI
Note: Wave 2 = 3 months post-release; Wave 3 = 9 month post-release; Wave 4 = 15 months post-release.
99
Multi-site Evaluation of SVORI: Summary & Synthesis
Adult SVORI program
participants were more
likely to report that they
were supporting
themselves with jobs.
Exhibit 47 shows the percentages of each group at each data
collection wave who reported that they were supporting
themselves with a job. Adult SVORI program participants were
more likely to report that they were supporting themselves with
jobs than the non-SVORI comparisons. Furthermore, the
percentages of adult SVORI program participants who reported
supporting themselves with jobs increased over time. The
differences in percentages between SVORI and non-SVORI
adults were statistically significant at the 15-month interview,
when 71% of men and 68% of women who had participated in
SVORI reported that they were supporting themselves with
jobs, compared with 60% and 45% of the non-SVORI men and
women.
Results for juveniles were less encouraging, although the
results were not statistically significant. Non-SVORI comparison
participants were more likely to report supporting themselves
with jobs than SVORI program participants at the 3- and 9month follow-up interview. There was a substantial increase
between the Wave 3 and Wave 4 interviews in the percentage
of SVORI participants who reported supporting themselves with
a job, so the percentage of this group surpassed that of the
non-SVORI comparison group, which also exhibited an increase.
Exhibit 48 shows the percentages who reported that their
current or most recent job provided formal pay. As can be
seen, most in all groups reported receiving formal pay, and
SVORI program participants were more likely to report formal
pay at all data collection waves. (Differences for the women at
all waves, for the men at Wave 1, and the juvenile males at
Wave 2 were statistically significant at alpha = 0.05.)
Interestingly, somewhat different patterns are seen for the
three demographic groups. The percentages of women
reporting formal pay remained approximately the same (by
group) across data collection waves, and SVORI program
participation was associated with a statistically significant
greater likelihood of reporting formal pay. However, the
percentages of adult male non-SVORI comparisons reporting
formal pay also remained roughly constant over time, whereas
the percentage of the adult male SVORI participants reporting
formal pay actually declined—from 84% 3 months after release
to 78% at 15 months after release.
100
Outcomes
Exhibit 47. Support of self with job by demographic group, study group, and data collection
wave
100%
80%
64%
60%
59%
53%
49%
32%
40%
40%
20%
0%
Wave 2
Men, SVORI
Men, non-SVORI
Women, SVORI
Women, non-SVORI
Juvenile, SVORI
Juvenile, non-SVORI
100%
80%
68%
68%
61%
56%
60%
32%
40%
39%
20%
0%
Wave 3
Men, SVORI
Men, non-SVORI
Women, SVORI
Women, non-SVORI
Juvenile, SVORI
Juvenile, non-SVORI
100%
80%
71%
60%
68%
60%
45%
53%
44%
40%
20%
0%
Wave 4
Men, SVORI
Men, non-SVORI
Women, SVORI
Women, non-SVORI
Juvenile, SVORI
Juvenile, non-SVORI
Note: Wave 2 = 3 months post-release; Wave 3 = 9 month post-release; Wave 4 = 15 months post-release.
101
Multi-site Evaluation of SVORI: Summary & Synthesis
Exhibit 48. Job with formal pay by demographic group, study group, and data collection
wave
100%
89%
84%
74%
80%
74%
70%
65%
60%
40%
20%
0%
Wave 2
Men, SVORI
Men, non-SVORI
Women, SVORI
Women, non-SVORI
Juvenile, SVORI
Juvenile, non-SVORI
90%
100%
80%
80%
83%
77%
73%
66%
60%
40%
20%
0%
Wave 3
Men, SVORI
Men, non-SVORI
Women, SVORI
Women, non-SVORI
Juvenile, SVORI
Juvenile, non-SVORI
90%
100%
80%
78%
74%
74%
76%
71%
60%
40%
20%
0%
Wave 4
Men, SVORI
Men, non-SVORI
Women, SVORI
Women, non-SVORI
Juvenile, SVORI
Juvenile, non-SVORI
Note: Wave 2 = 3 months post-release; Wave 3 = 9 month post-release; Wave 4 = 15 months post-release.
102
Outcomes
SVORI program
participants were more
likely to report that their
most recent job had
benefits.
SVORI programs appear
to have reduced
substance use among
program participants,
although overall drug use
increased over time for
all groups.
Exhibit 49 shows the percentage of each group who reported
that their current or most recent job had benefits (paid
vacation, health insurance, or both). Again, SVORI program
participation was associated with better outcomes (except for
the juvenile males 3 months after release). Adult males who
participated in SVORI were most likely to report having
benefits, and the difference between the SVORI and non-SVORI
respondents were statistically significant at all data collection
waves. SVORI program participation also appears to have
benefited the women who participated, because they were
more likely than their non-SVORI counterparts to report that
their jobs provided either paid vacation or health insurance.
Overall, however, the percentages were low—with only about
half (or less) reporting that their jobs provided health
insurance, paid leave, or both. Percentages were particularly
low for the women in the comparison group: Less than 40% at
any data collection wave reported having job benefits.
SVORI programs appear to have reduced substance use among
program participants, although overall drug use increased over
time for all groups and exceeded 50% at 15 months.22
Exhibit 50 shows a comprehensive abstinence measure
collected on substance use at Waves 2 and 4—no self-reported
use of any drug in the past 30 days and no positive results on
the oral swab drug test (a refusal was counted as a positive).
SVORI program participants performed better on this measure
at both waves, in some cases by a substantial margin. As can
be seen, 15 months following release, less than half were
abstinent with respect to substance use—ranging from a low of
38% of the non-SVORI adult female respondents to 46% of all
SVORI respondents.
22
Individuals incarcerated since the 9-month interview (or for the past
6 months if they did not complete a Wave 3 interview) were not
asked about substance use at 15 months.
103
Multi-site Evaluation of SVORI: Summary & Synthesis
Exhibit 49. Job with benefits by demographic group, study group, and data collection wave
100%
80%
60%
47%
40%
39%
41%
37%
26%
25%
20%
0%
Wave 2
Men, SVORI
Men, non-SVORI
Women, SVORI
Women, non-SVORI
Juvenile, SVORI
Juvenile, non-SVORI
100%
80%
60%
53%
42%
45%
41%
33%
40%
42%
20%
0%
Wave 3
Men, SVORI
Men, non-SVORI
Women, SVORI
Women, non-SVORI
Juvenile, SVORI
Juvenile, non-SVORI
100%
80%
60%
52%
44%
53%
42%
40%
37%
40%
20%
0%
Wave 4
Men, SVORI
Men, non-SVORI
Women, SVORI
Women, non-SVORI
Juvenile, SVORI
Juvenile, non-SVORI
Note: Wave 2 = 3 months post-release; Wave 3 = 9 month post-release; Wave 4 = 15 months post-release.
104
Outcomes
Exhibit 50. No self-reported drug use in the past 30 days or positive drug test by
demographic group, study group, and data collection wave
100%
80%
60%
67%
54%
62%
52%
49%
48%
40%
20%
0%
Wave 2
Men, SVORI
Men, non-SVORI
Women, SVORI
Women, non-SVORI
Juvenile, SVORI
Juvenile, non-SVORI
100%
80%
60%
46%
43%
46%
38%
40%
46%
43%
20%
0%
Wave 4
Men, SVORI
Men, non-SVORI
Women, SVORI
Women, non-SVORI
Juvenile, SVORI
Juvenile, non-SVORI
Note: Wave 2 = 3 months post-release; Wave 4 = 15 months post-release.
The levels of illicit substance use demand some consideration.
Three months after release, for SVORI program participants,
two thirds of the women and more than 60% of the juvenile
males were “clean” by the combined measure—but only about
half of the men. For the non-SVORI comparisons, about 50% of
all groups were “clean,” slightly fewer than the SVORI men. The
percentage clean at 15 months declined for all groups to less
than 50%. These results are not necessarily surprising, given
the high levels of reported use before incarceration.
At Wave 1, respondents were asked whether they used any
drugs in the 30 days preceding the current incarceration.
Between 66% and 70% of the adults, depending upon
demographic and study group, reported “yes,” and about 60%
of the juvenile males reported “yes” (data not shown). These
105
Multi-site Evaluation of SVORI: Summary & Synthesis
figures suggest no-use rates of 30% to 40%, which are worse
than the rates presented in Exhibit 50.
For example, 66% of adult male SVORI program participants
reported using drugs in the 30 days preceding incarceration, or
34% reported no use—which is 35% higher than the 46% rate
observed 15 months after release.23 Exhibit 51 shows the
“improvements” for all groups. What is shown is the percentage
difference between Wave 1 self-report of any drug use in the
30 days before incarceration and the Wave 4 combined
measure of no self-report and no positive test. So, despite the
high levels of use, all groups were somewhat better.
Exhibit 51 Differences in drug use in the past 30 days by demographic group, study group,
and data collection wave
100%
Men, SVORI
Men, non-SVORI
80%
Women, SVORI
Women, non-SVORI
Juvenile, SVORI
60%
Juvenile, non-SVORI
40%
35%
39%
35%
27%
20%
12%
10%
0%
Improvement in past-30-day drug use
Note: What is shown is the percentage difference between Wave 1 self-report of no drug use in the 30 days before
incarceration and the Wave 4 combined measure of no self-report and no positive test. Wave 1 = 30 days prerelease; Wave 2 = 3 months post-release; Wave 3 = 9 month post-release; Wave 4 = 15 months post-release.
23
106
There were no preincarceration drug test results. If test results had
been available, the levels of abstinence preincarceration likely
would have been lower (similar to the findings presented earlier for
post-release). Thus, the improvements from preincarceration to 15
months post-release presented here represent the least amount of
improvement.
Outcomes
SVORI program
participation appeared to
have little effect on
compliance with
conditions of supervision.
Men who participated in
SVORI programs were
more likely than their
comparison counterparts
to report having
committed no criminal
acts…results for the
women and juvenile
males were less
consistent.
Overall, SVORI program participation appeared to have little
effect on compliance with conditions of supervision. Exhibit 52
presents the results for all groups and data collection waves. As
can be seen, self-reported compliance declined over time,
particularly for the adults.
The men who participated in SVORI programs were more likely
than their comparison counterparts to report having committed
no criminal acts in the period since the most recent interview
(difference was significant at 3 months post-release).24
Exhibit 53 provides the data for each group and data collection
wave. As can be seen, the results for the women and juvenile
males were less consistent. At 3 and 9 months after release,
women who participated in SVORI were less likely to report
having committed no criminal activities than the adult female
comparison subjects, although this trend reversed at 15
months (differences not significant). For the juvenile males,
SVORI program participants were somewhat less likely to
report criminal behavior, but there were no differences between
the groups from the subsequent follow-up interviews.
The final measures of criminal activity to be considered here
were derived from administrative records—arrest records from
the NCIC and reincarceration derived from corrections data.
Results that follow are for adults only.
Although the SVORI men
had somewhat lower
arrest rates through the
24-month period, the
differences between
SVORI and non-SVORI
were not significant.
Exhibit 54 shows the cumulative post-release arrest rates for
the four adult groups over a 24-month follow-up period. What
is shown is the percentage of each group that has experienced
at least one new arrest by the indicated time after release. The
highest rates are recorded for the adult male comparisons—
with fully 71% having been rearrested within 2 years of
release. The lowest rates are recorded for the women who
participated in SVORI—49% had been rearrested within 2 years
of release, significantly fewer women than from the non-SVORI
comparison group. Although the SVORI men had somewhat
lower arrest rates through the 24-month period, the differences
between SVORI and non-SVORI were not significant.
24
The question about committing criminal activities was not asked of
those incarcerated for the entire reference period.
107
Multi-site Evaluation of SVORI: Summary & Synthesis
Exhibit 52. Self-reported compliance with supervision conditions by demographic group,
study group, and data collection wave
100%
80%
78%
78%
83%
78%
80%
86%
60%
40%
20%
0%
Wave 2
Men, SVORI
Men, non-SVORI
Women, SVORI
Women, non-SVORI
Juvenile, SVORI
Juvenile, non-SVORI
100%
80%
69%
76%
70%
71%
74%
59%
60%
40%
20%
0%
Wave 3
Men, SVORI
Men, non-SVORI
Women, SVORI
Women, non-SVORI
Juvenile, SVORI
Juvenile, non-SVORI
100%
80%
66%
60%
77%
68%
60%
57%
66%
40%
20%
0%
Wave 4
Men, SVORI
Men, non-SVORI
Women, SVORI
Women, non-SVORI
Juvenile, SVORI
Juvenile, non-SVORI
Note: Wave 2 = 3 months post-release; Wave 3 = 9 month post-release; Wave 4 = 15 months post-release.
108
Outcomes
Exhibit 53. Self-reported noncommission of criminal activity by demographic group, study
group, and data collection wave
100%
80%
79%
73%
82%
78%
53%
60%
48%
40%
20%
0%
Wave 2
Men, SVORI
Men, non-SVORI
Women, SVORI
Women, non-SVORI
Juvenile, SVORI
Juvenile, non-SVORI
100%
80%
64%
60%
75%
69%
59%
41%
41%
40%
20%
0%
Wave 3
Men, SVORI
Men, non-SVORI
Women, SVORI
Women, non-SVORI
Juvenile, SVORI
Juvenile, non-SVORI
100%
80%
66%
61%
70%
64%
60%
37%
40%
38%
20%
0%
Wave 4
Men, SVORI
Men, non-SVORI
Women, SVORI
Women, non-SVORI
Juvenile, SVORI
Juvenile, non-SVORI
Note: Wave 2 = 3 months post-release; Wave 3 = 9 month post-release; Wave 4 = 15 months post-release.
109
Multi-site Evaluation of SVORI: Summary & Synthesis
Exhibit 54. Cumulative arrest rates for adult SVORI and non-SVORI study participants
100%
Men, SVORI
Men, non-SVORI
Women, SVORI
Women, non-SVORI
80%
71%
68%
60%
60%
49%
40%
20%
0%
3
6
9
12
15
21
24
Months
There is little difference
in the reincarceration
rates of the SVORI men,
non-SVORI men, and the
SVORI women in terms of
their chances of being
reincarcerated within 24
months of release.
110
Exhibit 55 provides reincarceration information for the adult
study participants. These results look quite different from those
presented in Exhibit 54. In particular, there is little difference in
the reincarceration rates of the SVORI men, non-SVORI men,
and the SVORI women in terms of their chances of being
reincarcerated within 24 months of release—42% of the SVORI
men, 41% of the SVORI women, and 39% of the non-SVORI
men were reincarcerated after their release from the
incarceration that was associated with their inclusion in the
study. In sharp contrast, only 22% of the non-SVORI women
were reincarcerated. Explorations continue to identify possible
explanations for the differences between the self-report, arrest,
and reincarceration data. Potential factors include those
associated with supervision—numbers of conditions, revocation
practices (which likely vary across, if not within, state).
Outcomes
Exhibit 55. Cumulative reincarceration rates for adult SVORI and non-SVORI study
participants
100%
Men, SVORI
Men, non-SVORI
Women, SVORI
Women, non-SVORI
80%
60%
40%
42%
41%
39%
20%
22%
0%
3
6
9
12
15
21
24
Months
111
Economic
Evaluation
Decision makers who are in charge of planning and funding
prisoner reentry programs and policies need to know the cost of
reentry services and the benefits that accrue to society as a
result of these services. This part of the report summarizes the
results of an economic evaluation of the enhanced reentry
efforts funded through SVORI. It provides the results from two
separate analyses:
ƒ
Pre-release service cost analysis. A detailed cost
analysis was conducted of those services provided
before prisoners were released into the community in
2004 and 2005. Estimates are provided at the site level
and throughout the course of the 12 months
immediately preceding release.
ƒ
Cost-benefit analysis. A cost-benefit analysis (CBA) of
enhanced reentry was conducted, both pre- and postrelease. Monthly estimates are provided, and estimates
for all four adult sites are combined.
The next section provides an overview of the evaluation
methodology, including data sources, followed by a summary of
the findings. Readers interested in detailed information on the
methodology and findings are referred to Cowell et al. (2009).
METHODOLOGY OVERVIEW
The perspective of the analyses—which determines whose costs
and benefits are measured—is the criminal justice system.
State agencies in the criminal justice system underwrite all
services received by offenders (e.g., anger management) and
certainly pay for the law enforcement and court resources used
in rearresting and the prison resources used in reincarcerating
offenders.
113
Multi-site Evaluation of SVORI: Summary & Synthesis
The pre-release service cost analysis examined how resources
were spent on offenders before entering the community from
four adult sites—Iowa, Ohio, Pennsylvania, and South
Carolina—and one juvenile site, South Carolina. Pre-release
services are of particular interest to decision makers, such as
those in departments of corrections, because they are largely
within the direct control of grant recipients. After release,
services that offenders receive are typically provided by a
variety of providers, and decision makers often have less
control over provision and access. Estimates are for the 12
months before release, that is, annual costs.
The analysis provided
estimates on the costs per
SVORI participant over
and above a comparison
participant. These
estimates are referred to
as “incremental costs.”
The CBA addressed the
degree to which
expenditures on services
were offset by reductions
in the key outcomes of
rearrest and
reincarceration.
114
Because most corrections agencies understand, at a broad
level, the cost of standard reentry, the most practical
advantage of a cost analysis is to understand the additional
costs of enhanced reentry resources. Thus, in addition to the
cost of service provision per site, the analysis provided
estimates on the costs per SVORI participant over and above a
comparison participant. These estimates are referred to as
“incremental costs.” In addition to an estimated incremental
cost for the program as a whole, the results also describe the
incremental costs for specific domains of costs, such as
educational services. Disaggregating the estimates in this
manner is particularly useful because it reflects the way in
which appropriations and budgets are disaggregated and
recognizes that, beyond the SVORI grant, different components
of reentry programs in the community may have different
funding sources. Because of state-to-state variation, separate
analyses for each of the five programs selected for the cost
analysis were conducted.
The CBA addressed the degree to which expenditures on
services were offset by reductions in the key outcomes of
rearrest and reincarceration. It accounts for the fact that
services—before and after release—are designed to help
offenders contribute positively to society. The analysis
combined the data from the pre-release cost analysis with
estimates of post-release costs, which included costs incurred
by obtaining additional services and those incurred through
rearrest and reincarceration. The analysis then compared, at
key time points, average costs for two groups: those in
enhanced reentry and those in a comparison group. Relative to
the comparison group, the enhanced reentry group would likely
have higher service costs but lower arrest and incarceration
Economic Evaluation
costs, and total costs overall would likely be lower. For this
analysis, all adult sites’ data (the juvenile site data were
analyzed separately) were combined to increase sample size
and statistical power. It also was in keeping with the design of
the main outcome evaluation, which combined all sites in its
analyses.
SAMPLE
The sample consisted of 722 men from four adult sites and 79
juveniles from one juvenile male program. Because relatively
few women participated in services at these sites, women are
not included in the analyses. Each program had a pre-release
phase that almost exclusively provided services within the
correctional facility and a post-release phase that helped secure
services for offenders once they were in the community.
The programs were selected for the study by first conducting a
systematic screen of each evaluation site. Selection criteria
included the following:
ƒ
Sample size should be as large as possible. A common
challenge in economic evaluation is that program cost
estimates may vary greatly within the sample being
studied, making it difficult to find statistically significant
results. The difficulty stems from the fact that some
people consume relatively few resources, whereas
others are associated with relatively high expenditures.
Because the spread of cost estimates around the
measure of central tendency is large, the standard error
of the estimate is large, which means that statistical
significance may be difficult to achieve.
ƒ
The study designs should be as strong as possible. Two
sites in the study, Iowa and Ohio, randomized
participants into SVORI or a comparison group and
were, thus, included. In the case of a quasiexperimental design, a stronger design is one in which
the comparison group is very similar in key
characteristics to the SVORI sample. As documented
elsewhere (Lattimore et al., 2008), a careful selection
protocol for the comparison group at each site paid
dividends insofar as analyses indicated that the SVORI
and comparison samples matched well at baseline
across many measures.
ƒ
Availability of administrative records was also a practical
consideration. Sites were thus selected that had the
115
Multi-site Evaluation of SVORI: Summary & Synthesis
promise of providing collateral administrative data on
service utilization.
ƒ
At least one program selected should be targeted at
juveniles.
Five programs met at least one of these criteria. By
implementing random assignment to SVORI or comparison
study conditions, Iowa and Ohio had the strongest possible
design. The two other adult programs—South Carolina and
Pennsylvania—had both a sufficiently large sample size and a
sufficiently strong quasi-experimental sample design. The
South Carolina juvenile program had a relatively large number
of juveniles, could provide administrative data, and was able to
provide particularly detailed budgetary information for the
economic evaluation.
DATA
The data for analyses came from multiple sources. Because of
the variation in the types of services and the nature of the data
available, the data sources for service costs were particularly
diverse. Data came from a mixture of the offender surveys,
site-specific documentation, the literature, and expert opinion.
Data on the costs of arrests came from the Bureau of Justice
Statistics (BJS) and the literature, and data on incarceration
came from BJS. All dollar estimates were adjusted to 2007
costs using the Consumer Price Index of the Bureau of Labor
Statistics (BLS).
The offender survey was conducted as part of the main SVORI
evaluation and contained several questions that were used for
the two economic analyses. The survey instrument was quite
comprehensive and is described more fully in Lattimore and
Steffey (2009). Data on service use, arrest, and incarceration
were collected at four points in time: 1 month before release
and 3, 9, and 15 months after release. Because the survey data
only provide information on whether an offender received a
service, supplementary information was required on the
quantity of services received. The preferred source of
information on quantity was site-specific documentation,
obtained from on-site visits and existing program literature,
such as program descriptions, syllabi, or supporting materials
for contracts. Other supplementary sources included the
broader literature and opinions from site program staff and
116
Economic Evaluation
substantive experts. Sensitivity analyses (below) varied these
and other assumptions to assess the degree to which the study
conclusions depended on them. It also should be noted that
estimates of quantity and price were obtained only for those
services for which there was a statistically significant difference
in the proportion of SVORI and comparison group participants
reporting receipt at each site.
Information on the price for each service came from a
combination of program- and service-specific collateral sources
and the literature. Price data were obtained for services in
which there was a statistically significant difference between
the proportion of the SVORI and the comparison group
participants receiving the service.
Counts of the number of arrests and nights incarcerated came
from the offender surveys. The measure of arrest was the
number of times since the last interview that the respondent
had been booked into jail or prison, as based on responses at
each of the three follow-up interview waves. The measure of
nights incarcerated was the number of nights incarcerated since
the last interview, as indicated at each of the three follow-up
waves.25
The costs per arrest were constructed from estimates in the
literature. The estimates included eight major components of
costs, from initial contact, investigation, and arrest through
sentencing. Each individual component had a raw cost
associated with it. For example, on average (in 2007 dollars), it
costs $334 to book someone, and a trial costs, on average,
$55,088. Sensitivity analyses (the methods of which are
described below) used an alternative estimate of the cost of an
arrest to assess the degree to which conclusions were robust to
the specific arrest cost used.
RESULTS
Pre-release service costs varied considerably across the five
sites. Estimates of the additional pre-release cost attributable
to enhanced reentry varied from $658 per offender for the
Pennsylvania adult program to $3,480 for the South Carolina
juvenile program. The incremental costs were $1,163 for the
25
The data contained missing values for incarceration days that
should, in fact, be nonmissing, and these items were imputed. See
Cowell et al. (2009) for details.
117
Multi-site Evaluation of SVORI: Summary & Synthesis
Iowa program, $1,698 for the Ohio program, and $1,480 for
the South Carolina adult program. Case management
contributed largely to pre-release costs. The exception was in
Iowa, where the difference was driven by employment/
education/life skills services.
The CBA results for men provide estimates of monthly net costs
(services plus arrests plus incarceration) at each of the four
waves of the offender survey. The total net cost estimates
(bottom set of rows in the tables) indicate that, as expected,
there was a statistically significant net monthly cost over the 12
months before release (p < 0.05). At that point, enhanced
reentry through SVORI had $129 higher net costs per offender
per month. Almost all of this ($121 of the $129 difference) is
service programming. Offenders had almost the same
incarceration costs and no arrest costs in this period. None of
the differences in net costs at each of the three periods after
release are statistically significant at conventional levels. Three
months after release, the average monthly net cost of the two
groups was very close. The $6 cost savings for the SVORI
group was small. Nine months after release, enhanced reentry
was associated with $154 higher monthly costs (not statistically
significant; p > 0.10; 95% confidence interval [CI] = −79,387).
At 15 months after release, enhanced reentry was associated
with $248 higher monthly net costs, which is marginally
statistically significant (p < 0.10; 95% CI = −19,516).
The estimates indicate that, throughout the period of study,
enhanced reentry was associated with more resources being
used for services. The difference between the enhanced reentry
and comparison groups in monthly costs for services spiked at
the 3-month interview—at $216 (p < 0.05)—and then
diminished over time. At the 9-month follow-up interview, the
monthly cost difference was $105 (p < 0.05), and at the
15-month interview the difference was $97 (p < 0.05).
The difference between the enhanced reentry and comparison
groups in criminal justice costs was not statistically significant
at any of the three follow-up waves. The estimates indicate that
at the 3-month follow-up, the comparison group incurred $221
higher monthly criminal justice costs, whereas at 9 and 15
months, the enhanced reentry group incurred higher monthly
criminal justice costs at $49 and $152, respectively. However,
because of the lack of precision in the estimate, the sign and
118
Economic Evaluation
magnitude of any of these findings cannot be reliably
interpreted.
The results of the sensitivity analyses for men provide
minimum and maximum estimates around the main findings.
Overall, the sensitivity analyses do not change the conclusions
from the main findings. Service costs for the enhanced reentry
group were higher at every wave, with the difference peaking 3
months after release. As for the main analysis findings,
differences in criminal justice costs were not significantly
different at any of the waves. This holds for criminal justice
costs as a whole, as well as for its two components, arrests and
incarceration.
The main CBA findings for the South Carolina juvenile reentry
site showed that the differences in monthly service costs
between the SVORI enhanced reentry group and the
comparison were pronounced before release ($282; p < 0.05)
and 3 months after release ($330; p < 0.05). Thereafter,
service costs were similar for the two groups. There was
evidence that incarceration costs were higher for the juveniles
in enhanced reentry for at least one of the follow-up periods.
Average monthly incarceration costs for juveniles in the SVORI
group were higher than the comparison by $486 at 9 months
(p < 0.05; 95% CI = $47, $915) and by $859 at 15 months
(p < 0.05; 95% CI = $180, $1,537).
The number of observations in the juvenile data varied from 56
observations at 3 months post-release to 79 observations
before release. This number proved too low to yield reliable
results in additional multivariate analyses that control for
potential confounders (e.g., age). The low number of
observations also likely contributes to the large confidence
intervals around many of the estimates. The results of the
sensitivity analyses for the juvenile site did not change the
main conclusions with regard to either overall costs or each of
the three components of costs (i.e., services, arrests, and
incarceration).
CONCLUSIONS
The SVORI economic evaluation indicated that enhanced
reentry was successful in delivering services to offenders, both
before and after release into the community. The cost-benefit
analysis was inconclusive and could not determine whether net
119
Multi-site Evaluation of SVORI: Summary & Synthesis
costs for the SVORI group were higher or lower than the
comparison group.
A limitation of the findings is that the estimates were not
sufficiently precise to detect differences in overall criminal
justice costs at any point in time after release. Another
limitation is that detailed information on the number and the
specific type of service events was not available. A final
limitation is that results may not generalize to other
jurisdictions.
Given the richness of the SVORI evaluation data, further work
assessing the impact on criminal justice costs may be
rewarding. These future directions will help address gaps in
current knowledge. First, future work is needed to further
examine the degree to which enhanced reentry programming
may be associated with possible reductions in criminal justice
costs. This may be perhaps most successfully done by using
administrative rather than survey data on arrests and
reincarceration and expanding the analysis to all of the 16 sites
used in the main evaluation for which reliable data are
available. A second likely fruitful area is to expand the kinds of
benefits examined to include outcomes such as employment.
120
Conclusions, Policy
Implications, and
Future Work
The emerging consensus that the complexity of disadvantages
confronting adults and juveniles after their release from prisons
and detention facilities points to a need for comprehensive
reentry strategies provided the context for the federal SVORI.
The $500 thousand to $2 million, 3-year grants funded through
SVORI were used to establish programs to develop, enhance, or
expand programs to facilitate the reentry of adult and juvenile
offenders returning to communities from prisons or juvenile
detention facilities. These programs were to span the periods
before release, in the community on supervision, and postsupervision.
The findings in this report have provided information on the
characteristics and experiences of adult male and female and
juvenile male SVORI program participants and comparison
subjects in 14 states who were released from prison or juvenile
detention between July 2004 and November 2005. Many of the
services provided by the programs were intended to improve
intermediate outcomes that have been correlated with
recidivism—for example, employment services to improve
employment, substance use treatment to reduce use, and
cognitive programs to address criminal thinking associated with
crime. The underlying model suggests that improvements in
these outcomes will lead to reductions in criminal behavior. The
SVORI program participants (and, to a lesser extent, the nonSVORI respondents) received various services, each of which
may have impacted one or more intermediate outcomes that, in
turn, may have impacted recidivism. There is little theoretical
121
Multi-site Evaluation of SVORI: Summary & Synthesis
or empirical guidance for the correct specification of such a
complex recidivism model, so the approach to the outcome
analyses was to test first-order effects of SVORI program
participation on each of the identified outcomes, including
recidivism.
The findings in this summary and synthesis report and the full
reports (Cowell et al., 2009; Hawkins et al., 2009; Lattimore &
Steffey, 2009; Lattimore et al., 2008; Lindquist et al., 2009)
provide information on the characteristics and experiences of
SVORI program participants and comparison subjects in 14
states who were released from prison between July 2004 and
November 2005. Sixteen (12 adult and 4 juvenile) of the 89
SVORI programs were included in the impact evaluation.
The findings substantiate that prisoners returning to their
communities are a population with extremely high needs and
that their expressed needs diminished somewhat, but remained
high, up to 15 months after release from prison or detention.
Overall, the participants in the study had weak educational and
employment histories, extensive substance use histories,
substantial experience with the criminal justice system, and
extensive exposure to drug or criminally involved family
members and peers. In particular, nearly all had used drugs in
the past, and most had used drugs in the 30 days before their
incarceration or detention. Most had been previously
incarcerated or, for juvenile participants, detained in juvenile
facilities. A majority had been treated for mental health or
substance use problems. Most acknowledged their need for
services, programming, and support, with substantial
proportions indicating need for services—particularly services
related to employment/education/skills and their transition to
the community, and, for the women, health. Although
expressed need diminished between the first interview
conducted about 30 days before release and the first follow-up
interview conducted about 3 months after release, the decline
was not substantial and remained relatively stable over the
next year for most service items. Expressed need for services
was similar for SVORI participants and non-SVORI respondents.
122
Conclusions, Policy Implications, and Future Work
SVORI program
participation increased
the likelihood of adult
participants’ receiving a
wide range of services.
SVORI programs were
unable to sustain levels
of service provision to
respondents with high
levels of expressed
needs after release.
Results from the impact study suggest that adult SVORI
programs were successful in significantly increasing access to a
variety of services and programming—particularly services
related to transitioning to the community and employment/
education/skills, as well as to substance use treatment. For
example, 75% of adult male SVORI program participants, in
contrast to 51% of non-SVORI comparison subjects, reported
involvement while in prison in programs to prepare for release.
This approximately 50% increase in the likelihood of reentry
program participation that was observed for the SVORI
program participants was seen more broadly across a range of
services and in the bundle scores for both adult males and
females. These findings add to emerging research regarding the
feasibility of improving service receipt across broadly conceived
reentry programs.
Although the adult SVORI programs were successful in
increasing the types and amounts of needs-related services
provided before and after release from prison, the proportion of
individuals who reported receiving services was smaller than
the proportion reporting need (sometimes much smaller) and,
generally, smaller than the SVORI program directors expected.
This finding is consistent with the fact that SVORI sites were
still developing and implementing their programs; it is a
reminder that complex programs may require a sustained effort
over several years to achieve full implementation.
Service delivery to adults declined substantially, on average,
after release—although somewhat less so for the adult females.
As noted, there was also a decline in expressed need for
services, but the decline in reported service receipt was
substantially larger. Thus, overall, the programs were unable to
sustain support for individuals during the critical, high-risk
period immediately after release. This decline may be
attributable to the difficulty programs experienced early in their
efforts to identify and coordinate services for individuals
released across wide geographic areas and, again, suggests the
need for sustained effort to achieve full implementation.
There were fewer differences in services provided to juveniles
in the four SVORI programs included in the impact evaluation
as compared with their non-SVORI counterparts, although
levels of services were generally higher than those observed for
the adult males. In general, women participating in SVORI
123
Multi-site Evaluation of SVORI: Summary & Synthesis
programs reported receiving more services than any other
group.
Participation in SVORI
programs was associated
with moderately better
outcome…for adults.
Findings were more
mixed for juveniles.
The significant—though less-than-universal—increase in service
receipt associated with participation in SVORI programs was
associated with moderately better outcomes with respect to
housing, employment, substance use, and self-reported
criminal behavior for adults. Findings were more mixed for
juveniles, for whom, as noted previously, there were fewer
differences in the types of services provided.
The SVORI logic model suggests that services responsive to
needs will result in improvements in intermediate outcomes,
including housing, employment, and substance use.
Improvements in these outcomes, in turn, are hypothesized to
result in improvements in criminal behavior. Consequently, for
most services, the service is linked to criminal behavior through
another, more intermediate outcome. (The exception is
programming to address criminal thinking, addressed below.)
Thus, education and employment-related services improve
employment outcomes, and a better job should reduce the
incentives to engage in criminal behavior. Substance use
treatment reduces drug use, which in turn should reduce the
need to commit property crimes (to pay for drugs) or violent
crimes that result from a lack of self-control associated with
substance use. Recidivism is a distal outcome linked to services
and programming through a set of intermediate outcomes.
The recidivism results were mixed. Adult male SVORI program
participants were less likely to report criminal activity
throughout the follow-up period (significantly so at 3 months
post-release, when 79% of SVORI and 73% of non-SVORI
reported no crimes since release). The differences remained at
about 6 percentage points over successive time points (64%,
as opposed to 59% at 9 months; 66%, as opposed to 61% at
15 months), implying that about 10% more SVORI program
participants reported no crimes during the previous 6 months
than non-SVORI comparisons. Adult female SVORI participants
were more likely than non-SVORI comparisons to report
criminal activity at Waves 2 and 3, but less likely at wave 4
(70% SVORI, as opposed to non-SVORI). The juvenile males in
SVORI programs were more likely 3 months after release to
report no criminal activity than the non-SVORI respondents,
124
Conclusions, Policy Implications, and Future Work
but the differences at the later data collection points were
negligible.
The participants in juvenile SVORI programs received more
services than non-SVORI comparison respondents—but,
generally, not significantly more. Few differences were
observed in outcomes, although SVORI program participants
were more likely to have no drug use 9 and 15 months after
release.
Rearrest rates, overall,
were high…by 24 months
after release, about 70%
of the men, 60% of the
non-SVORI women, and
49% of the SVORI women
had a new arrest.
Although self-reported
criminal behavior and
official arrest records
were consistent in
supporting somewhat
lower criminal activity
among SVORI program
participants, this pattern
was not associated with
lower reincarceration
rates.
Administrative arrest data were available only for the adult
respondents. Cumulative rearrest rates were calculated for 3,
6, 9, 12, 15, 21, and 24 months after release. SVORI program
participants were less likely to have an officially recorded
rearrest during the 24-month period after release. The
differences were small and not significant for the men but were
substantial and significant beginning with the 9-month data
point for the women (differences of 10 percentage points or
more). Rearrest rates, overall, were high for both groups—by
24 months after release, about 70% of the men, 60% of the
non-SVORI women, and 49% of the SVORI women had a new
arrest recorded at NCIC.
Although self-reported criminal behavior and official arrest
records were consistent in supporting somewhat lower criminal
activity among SVORI program participants, this pattern was
not associated with lower reincarceration rates. By 24 months
post-release, the reincarceration rate for adult male SVORI
program participants was about 8% higher than the non-SVORI
rate (42%, as opposed to 39%). For the women, 41% of SVORI
program participants had been reincarcerated by 24 months
after release, compared with only 22% of non-SVORI
counterparts.
Explanations for the reincarceration findings are the subject of
ongoing investigation. Among the potential hypotheses are the
following:
ƒ
The classic null hypothesis. Prisoner reentry
programs that provide services addressing needs such
as employment, education, and substance use treatment
do not affect criminal behavior even if they have
significant effects on the targeted outcomes (e.g., better
employment associated with a reentry program that
includes employment-related services).
125
Multi-site Evaluation of SVORI: Summary & Synthesis
ƒ
Insufficient statistical power associated with the
incomplete implementation of the reentry
programs. This explanation derives particularly from
the very low levels of services delivered—on average—
after release. This issue may be a particularly acute one
for the women and juvenile males, for whom total
sample sizes were small; but, even with the adult male
sample, the differences between treatments provided to
the two groups may have been insufficient, given level
of need.
ƒ
The impact of supervision practices. Adult SVORI
program participants were somewhat more likely than
non-SVORI adults to be on supervision after release,
placing them at greater risk of revocation.
Administrative data from the 12 states that were home
to adult programs were insufficient in some cases to
distinguish a readmission due to a technical revocation
from one associated with a new crime. The results for
the women are particularly difficult to explain—because
the administrative data suggest that SVORI program
participants are much less likely to be arrested but much
more likely to be reincarcerated. Many of the
comparison women were from Indiana. If Indiana’s
supervision practices differ substantially from the other
states—in particular, if Indiana was much less likely to
revoke—this difference may have generated a
confounder in the data. Although, with a propensity
score model, good balance was obtained for the SVORI
and non-SVORI groups, site was not included in the
model. Some preliminary analyses have been conducted
with the data from the adult females, but these have not
shown an “Indiana effect.” The results for the adult
females remain puzzling and warrant continued
investigation.
ƒ
The impact of averaging. All analyses that have been
presented have been for the total samples (i.e., men in
12 states, women in 11 states, and juvenile males in 4
states). It is known that substantial variation existed in
the level of services provided in these sites—both to
SVORI program participants and as the “status quo.”
Additional work is required to assess the relationship
between services provided and outcomes.
From a policy perspective, the multi-site SVORI evaluation adds
to the sparse reentry evaluation literature that addresses the
effect of broad-based (wraparound) programmatic efforts on
high risk individuals. Specifically, much of the reentry literature
to date presents findings from single-focus interventions, such
126
Conclusions, Policy Implications, and Future Work
as drug treatment or cognitive behavior therapy, which have
been implemented with low-risk offenders. SVORI was initiated
as consensus began to build that programs needed to address
the multiplicity of needs of offenders and that interventions
were likely to be more successful when focused on high-risk (or
higher risk) offenders.
It is important to remember that these programs were
evaluated during their first years of implementation. Most of
the SVORI impact programs were deployed in multiple prisons
and enrolled participants who returned to multiple
communities. Developing and implementing the panoply of
services for a comprehensive reentry program within multiple
prisons and identifying and enlisting community programs and
resources are complex tasks that could easily take several
years to fully realize. Thus, for example, although “only” 57%
of adult male SVORI participants reported having a reentry plan
30 days before release; this rate is 138% of the rate of nonSVORI respondents who reported having a plan. The 57%
finding suggests an opportunity for continued program
improvement and more complete implementation. Indeed,
many states appear to have viewed their program development
and implementation with SVORI funds as a foundation on which
to build better programs—by enhancing services and expanding
the reach of the services. As reported in Winterfield, Lindquist,
and Brumbaugh (2007), most SVORI program directors said in
response to a 2006 survey that their states were continuing to
build on the programs that they established with SVORI grant
funds.
Importantly, service delivery was not sustained during the
critical, high-risk period immediately after release. The
treatment literature suggests that 90 to 270 days of continued
care is optimal (Friedmann, Taxman, & Henderson, 2007;
Taxman, Perdoni, & Harrison, 2007). Larger program effects
may have accompanied continuous service delivery after
release.
The modest improvements in intermediate outcomes observed
in the evaluation of SVORI are consistent with findings from
several meta-analyses of single-program efforts. These
analyses suggest treatment effects from 10% to 20% across a
wide range of types of programming for offenders (Aos et al.,
2006). Whether a multifocus reentry program can lead to
127
Multi-site Evaluation of SVORI: Summary & Synthesis
significantly greater treatment effects of 30% to 50% is
unknown. Results from the SVORI evaluation suggest that sites
will need to be given sufficient time to implement
multicomponent, multiphase programs before this hypothesis
can be tested.
The evaluation was designed to address the question of
whether SVORI programs—enhanced reentry programs—can
impact the post-release outcomes of high-risk offenders. In
other words, the goal was to answer the question, “Did SVORI
work?” SVORI programs were “black boxes” that under the
SVORI model were assumed to contain the need-based services
appropriate for each individual. Although the programs differed
across sites, all programs conformed to this higher-order
definition of program. Deficiencies in service delivery are,
therefore, ascribed to development or implementation
shortcomings. Of course, this explanation is not wholly
satisfactory. Indeed, the extensive data collection on service
receipt was intended to allow examination of other evaluation
questions that were beyond the scope of this evaluation.
Additional evaluation questions remain—answers to which may
help guide policy—that were not addressed by this SVORI
evaluation, which was intended to determine whether SVORI
“worked.” These questions address “what worked” (Were some
SVORI programs more successful than others? Can the effects
of specific program components be disentangled?); “for whom”
(Are there identifiable characteristics that are associated with
better outcomes?); “for how long” (How are study participants
faring 5 years after release from prison?); and “at what cost?”
(Are long-term cost savings associated with the SVORI
programs?).
Additional research is required in order to determine the
answers to these questions. This research can address
questions generated by the hypotheses posed here. For
example, one related hypothesis to be tested involves the
question whether services directed at proximal outcomes (e.g.,
substance use treatment) are sufficient to effect changes in
criminal behavior, or whether programs targeted directly at
changing criminal thinking may be necessary, as well.
Additional examination of site-level differences is also
necessary to determine whether the characteristics of the sites
(e.g., parole revocation policies, economic climate) have an
128
Conclusions, Policy Implications, and Future Work
impact on “what works,” independent of local program and
participant characteristics. The extensive SVORI data set
provides an opportunity for future research to explore these
questions, as well as related questions, such as which services
were helpful, what factors led to reincarceration, and what
factors were associated with remaining out of prison.
129
References
Abrams, L. S. (2006). Listening to juvenile offenders: Can
residential treatment prevent recidivism? Child and
Adolescent Social Work Journal, 23(1), 61-85.
Abrams, L. S., Shannon, S., & Sangalang, C. (2008). Transition
services for incarcerated youth: A mixed methods
evaluation study. Children and Youth Services Review,
30, 522-535.
Allison, P. D. (2001). Missing data. Sage university papers
series on quantitative applications in the social sciences,
No. 07-136. Thousand Oaks, CA: Sage Publications.
Altschuler, D. M., & Brash, R. (2004). Adolescent and teenage
offenders confronting the challenges and opportunities
of reentry. Youth Violence and Juvenile Justice, 2(1), 7287.
Altschulter, D. M., & Armstrong, T. L. (1994). Intensive
aftercare for high risk juveniles: A community care
model. Washington, DC: Office of Juvenile Justice and
Delinquency Prevention.
Andrews, D. A. (2006). Enhancing adherence to risk-needresponsivity: Making quality a matter of policy.
Criminology and Public Policy, 5, 595-602.
Andrews, D. A., & Bonta, J. (2006). The psychology of criminal
conduct (4th ed.). Newark, NJ: Lexis/Nexis/Matthew
Bender.
Andrews, D. A., Ziner, I., Hoge, R. D., Bonta, J., Gendreau, P.,
& Cullen, F. T. (1990). Does correctional treatment
work: A clinically-relevant and psychologically-informed
meta-analysis. Criminology, 28, 369-404.
Aos, S., Miller, M., & Drake, E. (2006). Evidence-based adult
corrections programs: What works and what does not.
Olympia, WA: Washington State Institute for Public
Policy.
130
References
Bonczar, T. P., & Beck, A. J. (1997). Lifetime likelihood of going
to state or federal prison. Bureau of Justice Statistics
Special Report No. NCJ 1600092. Washington, DC: U.S.
Bureau of Justice Statistics.
Bouffard, J. A., & Bergseth, K. J. (2008). The impact of reentry
services on juvenile offenders’ recidivism. Youth
Violence and Juvenile Justice, 6, 295-318.
Chung, H., Schubert, C. A., & Mulvey, E. P. (2007). An
empirical portrait of community reentry among serious
juvenile offenders in two metropolitan cities. Criminal
Justice and Behavior, 34, 1402-1426.
Clear, T. R., Rose, D. R., & Ryder, J. A. (2001). Incarceration
and the community: The problem of removing and
returning offenders. Crime & Delinquency, 47, 335-351.
Cocozza, J. J., & Skowya, K. (2000). Youth with mental health
disorders: Issues and emerging responses. Juvenile
Justice, 7, 3-13.
Cowell, A., Roman, J., & Lattimore, P. K. (2009). An economic
evaluation of the Serious and Violent Offender Reentry
Initiative. Research Triangle Park, NC: RTI International.
Cullen, F. T., & Gendreau, P. (2000). Assessing correctional
rehabilitation: Policy, practice, and prospects. In J.
Horney (Ed.), Criminal justice 2000: Policies, processes,
and decisions of the criminal justice system (Vol. 3, pp.
109-175). Washington, DC: National Institute of Justice.
D’Agostino, R. B., Jr. (1998). Tutorial in biostatistics:
Propensity score methods for bias reduction in the
comparison of a treatment to a non-randomized control
group. Statistics in Medicine, 17, 2265-2281.
Farrington, D. P. (1989). Early predictors of adolescent
aggression and adult violence. Violence and Victims,
4(2), 79-100.
Feld, B. C. (1998). Juvenile and criminal justice systems’
responses to youth violence. In M. Tonry & M. H. Moore
(Eds.), Youth violence (pp. 189-263). Chicago:
University of Chicago Press.
Freudenberg, N., Daniels, J., Crum, M., Perkins, T., & Richie, B.
(2005). Coming home from jail: The social and health
consequences of community reentry for women, male
adolescents, and their families and communities.
American Journal of Public Health, 95(10), 1725-1736.
131
Multi-site Evaluation of SVORI: Summary & Synthesis
Friedmann, P. D., Taxman, F. S., & Henderson, C. (2007).
Evidence-based treatment practices for drug-involved
adults in the criminal justice system. Journal of
Substance Abuse Treatment, 32(3), 267-277.
Gaes, G. G., Flanagan, T. J., Motiuk, L. L., & Stewart, L.
(1999). Adult correctional treatment. In M. Tonry & J.
Petersilia (Eds.), Prisons (pp. 361-426). Chicago:
University of Chicago Press.
Glaze, L. E., & Palla, S. (2005). Probation and parole in the
United States, 2004. Washington, DC: U.S. Bureau of
Justice Statistics.
Glueck, S., & Glueck, E. (1974). Of delinquency and crime.
Springfield, IL: Thomas.
Glynn, R. J., Schneeweiss, S., & Sturmer, T. (2006). Indications
for propensity scores and review of their use in
pharmacoepidemiology. Basic Clinical Pharmacology and
Toxicology, 98(3), 253-259.
Gottfredson, M. R., & Hirschi, T. (1990). A general theory of
crime. Stanford, CA: Stanford University Press.
Grisso, T. (2004). Double jeopardy: Adolescent offenders with
mental disorders. Chicago: University of Chicago Press.
Hagan, J., & Coleman, J. P. (2001). Returning captives of the
American war on drugs: Issues of community and family
reentry. Crime & Delinquency, 47(3), 352-267.
Hagan, J., & Dinovitzer, R. (1999). Collateral consequences of
imprisonment for children, communities, and prisoners.
In J. Petersilia & M. Tonry (Eds.), Prisons (pp. 121-162).
Chicago: University of Chicago Press.
Harrison, P. (2000). Total sentenced prisoners released from
state or federal jurisdiction, National Prisoners Data
Series (NPS-1). Washington, DC: U.S. Bureau of Justice
Statistics. Available at:
http://www.ojp.usdoj.gov/bjs/data/corpop22.csv.
Harrison, P. M., & Beck, A. J. (2005). Prison and jail inmates at
midyear 2004 (NCJ 195189). Washington, DC: U.S.
Bureau of Justice Statistics.
Hawkins, J. D., Herrenkohl, T., Farrington, D. P., Brewer, D.,
Catalano, R. F., & Harachi, T. W. (1998). A review of
predictors of youth violence. In R. Loeber & D. P.
Farrington (Eds.), Serious and violent juvenile offenders:
Risk factors and successful interventions (pp. 106-146).
Thousand Oaks, CA: Sage Publications.
132
References
Hawkins, S., Dawes, D., Lattimore, P. K., & Visher, C. A.
(2009). Detainee reentry experiences of juvenile males:
Characteristics, service receipt, and outcomes of
participants in the SVORI Multi-site Evaluation. Research
Triangle Park, NC: RTI International.
Kazdin, A. (2000). Adolescent development, mental disorders,
and decision making in delinquent youth. In T. Grisso &
R. Schwartz (Eds.), Youth on trial (pp. 33-65). Chicago:
University of Chicago Press.
Lattimore, P. K. (2007). The challenges of reentry. Corrections
Today, 69(2), 88-91.
Lattimore, P. K., Brumbaugh, S., Visher, C., Lindquist, C. H.,
Winterfield, L., Salas, M., et al. (2004). National portrait
of the Serious and Violent Offender Reentry Initiative.
Research Triangle Park, NC: RTI International.
Lattimore, P. K., & Steffey, D. M. (2009). The Multi-Site
Evaluation of SVORI: Methodology and analytic
approach. Research Triangle Park, NC: RTI
International.
Lattimore, P. K., Steffey, D. M., & Visher, C. A. (2009). Prisoner
reentry experiences of adult males: Characteristics,
service receipt, and outcomes of participants in the
SVORI Multi-site Evaluation. Research Triangle Park, NC:
RTI International.
Lattimore, P. K., Visher, C. A., & Steffey, D. M. (2008). Prerelease characteristics and service receipt among adult
male participants in the SVORI Multi-site Evaluation.
Research Triangle Park, NC: RTI International.
Lattimore, P. K., Visher, C. A., Winterfield, L., Lindquist, C., &
Brumbaugh, S. (2005). Implementation of prisoner
reentry programs: Findings from the Serious and Violent
Offender Reentry Initiative Multi-site Evaluation. Justice
Research and Policy, 7(2), 87-109.
Laub, J. H., & Sampson, R. J. (1993). Turning points in the life
course: Why change matters to the study of crime.
Criminology, 31, 301-326.
Laub, J. H., & Sampson, R. J. (2001). Understanding desistance
from crime. In M. H. Tonry (Ed.), Crime and Justice: A
Review of Research. Chicago: University of Chicago
Press.
Laub, J. H., & Sampson, R. J. (2003). Shared beginnings,
divergent lives: Delinquent boys to age 70. Cambridge,
MA: Harvard University Press.
133
Multi-site Evaluation of SVORI: Summary & Synthesis
Lindquist, C. (2005). Reentry research in action:
Implementation of SVORI programs. Research Triangle
Park, NC: RTI International.
Lindquist, C., & Winterfield, L. (2005). Reentry research in
action: Characteristics of prisoner reentry programs.
Research Triangle Park, NC: RTI International.
Lindquist, C. H., Barrick, K., Lattimore, P. K., & Visher, C. A.
(2009). Prisoner reentry experiences of adult females:
Characteristics, service receipt, and outcomes of
participants in the SVORI Multi-site Evaluation. Research
Triangle Park, NC: RTI International.
Lipsey, M. W. (2009). The primary factors that characterize
effective interventions with juvenile offenders: A metaanalytic overview. Victims & Offenders, 4(2), 124–147.
Lipsey, M. W., & Cullen, F. T. (2007). The effectiveness of
correctional rehabilitation: A review of systematic
reviews. Annual Review of Law and Social Science, 3,
297-320.
Lipsey, M. W., & Derzon, J. H. (1998). Predictors of violent or
serious delinquency in adolescence and early adulthood:
A synthesis of longitudinal research. In R. Loeber & D. P.
Farrington (Eds.), Serious and violent juvenile offenders:
Risk factors and successful interventions (pp. 86-105).
Thousand Oaks, CA: Sage Publications.
Lowenkamp, C. T., Latessa, E. J., & Smith, P. (2006). Does
correctional program quality really matter? The impact
of adhering to the principles of effective intervention.
Criminology and Public Policy, 5, 575-608.
Lynch, J., & Sabol, W. (2001). Prisoner reentry in perspective.
Washington, DC: Urban Institute.
Mackenzie, D. L. (2006). What works in corrections: Reducing
the criminal activities of offenders and delinquents. New
York: Cambridge University Press.
Marlowe, D. B. (2006). When “what works” never did: Dodging
the “scarlet M” in correctional rehabilitation. Criminology
and Public Policy, 5(2), 339-346.
Maruna, S. (2001). Making good: How ex-offenders reform and
reclaim their lives. Washington, DC: American
Psychological Association Books.
Mauer, M. (2000). Race to incarcerate. Washington, DC: The
Prison Project.
134
References
McCord, J. (1992). The Cambridge-Somerville study: A
pioneering longitudinal-experimental study of
delinquency prevention. In J. McCord & R. Tremblay
(Eds.), Serious and violent juvenile offenders: Risk
factors and successful interventions (pp. 313-345).
Thousand Oaks, CA: Sage Publications.
Mears, D. (2001). Critical challenges in addressing the mental
health needs of juvenile offenders. Justice Policy Journal,
4, 41-61.
Mears, D. P., & Travis, J. (2004). Youth development and
reentry. Youth Violence and Juvenile Justice, 2(1), 3-20.
National Council on Disability (2003). Addressing the needs of
youth with disabilities in the juvenile justice system: The
current status of evidence-based research. Washington,
DC: National Council on Disability.
National Research Council (2007). Parole, desistance from
crime, and community integration. Washington, DC:
National Academy Press.
National Research Council and Institute of Medicine (2001).
Juvenile crime, juvenile justice. Washington, DC:
National Academy Press.
Petersilia, J. (2003). When prisoners come home: Parole and
prisoner reentry. New York: Oxford University Press.
Petersilia, J. (2004). What works in prisoner reentry? Reviewing
and questioning the evidence. Federal Probation, 68(2),
4-8.
Petersilia, J. (2005). Hard time: Ex-offenders returning home
after prison. Corrections Today, 155, 66-71.
Puzzanchera, C. (2009). Juvenile arrests 2007 (NCJ 225344).
Office of Juvenile Justice and Delinquency Prevention
Juvenile Justice Bulletin. Retrieved June 11, 2009, from
http://www.ncjrs.gov/pdffiles1/ojjdp/225344.pdf
Re-entry Policy Council (2005). Report of the Re-entry Policy
Council: Charting the safe and successful return of
prisoners to the community. New York: Council of State
Governments.
Rhine, E., Mawhorr, T., & Parks, E. C. (2006). Implementation:
The bane of effective correctional programs. Criminology
and Public Policy, 5(2), 347-358.
135
Multi-site Evaluation of SVORI: Summary & Synthesis
Rosenbaum, P., & Rubin, D. B. (1983). The central role of the
propensity score in observational studies for causal
effects. Biometrika, 70, 41-55.
Rosenbaum, P., & Rubin, D. B. (1985). The bias due to
incomplete matching. Biometrics, 41, 103-116.
Rossman, S. B., & Roman, C. G. (2003). Case-managed reentry
and employment: Lessons from the Opportunity to
Succeed Program. Justice Research and Policy, 5(2), 75100.
Rubin, D. B. (1997). Estimating causal effects from large data
sets using propensity scores. Annals of Internal
Medicine, 127, 757-763.
Rubin, D. B. (2006). Matched sampling for causal effects. New
York: Cambridge University Press.
Sampson, R. J., & Laub, J. H. (1993). Crime in the making:
Pathways and turning points through life. Cambridge,
MA: Harvard University Press.
SAS Institute (2004). SAS/STAT 9.1 user’s guide. Cary, NC:
Author.
Seiter, R. P., & Kadela, K. R. (2003). Prisoner reentry: What
works, what does not, and what is promising. Crime &
Delinquency, 49(3), 360-388.
Sickmund, M., Sladky, T. J., & Kang, W. (2008). Census of
juveniles in residential placement databook. Retrieved
December 18, 2009, from
http://www.ojjdp.ncjrs.gov/ojstatbb/cjrp/
Snyder, H. N., & Sickmund, M. (2006). Juvenile offenders and
victims: 2006 national report. Washington, DC: Office of
Juvenile Justice and Delinquency Prevention.
Steinberg, L., Chung, H. L., & Little, M. (2004). Reentry of
young offenders from the justice system: A
developmental perspective. Youth Violence and Juvenile
Justice, 2, 21-38.
Strategic Advantages (2000). Symptom Assessment-45
Questionnaire (SA-45). New York: Multi-Health Systems.
Taxman, F. S., Perdoni, M., & Harrison, L. (2007). Treatment
for adult offenders: A review of the state of the state.
Journal of Substance Abuse Treatment, 32(3), 239-254.
136
References
Travis, J. (2005). But they all come back: Facing the challenges
of prisoner reentry. Washington, DC: Urban Institute
Press.
Visher, C. A. (2006). Effective reentry programs [Editorial
introduction]. Criminology and Public Policy, 5(2), 299302.
Visher, C. A. (2007). Returning home: Emerging findings and
policy lessons about prisoner reentry. Federal
Sentencing Reporter, 20(2), 93-102.
Visher, C. A., & Travis, J. (2003). Transitions from prison to
community: Understanding individual pathways. Annual
Review of Sociology, 29, 89-113.
Ware, J., Jr., Kosinski, M., & Keller, S. D. (1996). A 12-Item
Short-Form Health Survey: Construction of scales and
preliminary tests of reliability and validity. Medical Care,
34(3), 220-233.
Warr, M. (1998). Life-course transitions and desistance from
crime. Criminology, 36, 183–216.
West, H. C., Sabol, W. J., & Cooper, M. (2009). Prisoners in
2008 (NCJ-228417). U.S. Bureau of Justice Statistics.
Retrieved December 27, 2009, from
http://bjs.ojp.usdoj.gov/index.cfm?ty=pbdetail&iid=1763.
Western, B. (2007). The penal system and the labor market. In
S. Bushway, M. A. Stoll, & D. F. Weiman (Eds.), Barriers
to re-entry? The labor market for released prisoners in
postindustrial America. New York: Russell Sage
Foundation.
Wilson, J. A., & Davis, R. C. (2006). Good intentions meet hard
realities: An evaluation of the project greenlight reentry
program. Criminology and Public Policy, 5(2), 303-338.
Winterfield, L., & Brumbaugh, S. (2005). Reentry research in
action: Overview of the Serious and Violent Offender
Reentry Initiative. Research Triangle Park, NC: RTI
International.
Winterfield, L., Lattimore, P. K., Steffey, D. M., Brumbaugh, S.,
& Lindquist, C. (2006). The Serious and Violent Offender
Reentry Initiative: Measuring the effects on service
delivery. Western Criminology Review, 7(2), 3-19.
Winterfield, L., & Lindquist, C. (2005). Reentry research in
action: Characteristics of prisoner reentry programs.
Research Triangle Park, NC: RTI International.
137
Multi-site Evaluation of SVORI: Summary & Synthesis
Winterfield, L., Lindquist, C., & Brumbaugh, S. (2007).
Sustaining adult reentry programs after SVORI. The
Multi-site Evaluation of the Serious and Violent Offender
Reentry Initiative: Reentry research in action. Retrieved
December 18, 2009, from http://www.svorievaluation.org/%5Cdocuments%5Creports%5CRRIA_2007_Ad
ult_Sustainability.pdf
Young, D. (2004). First count to ten: Innovation and
implementation to juvenile reintegration programs.
Retrieved December 15, 2009, from http://www.uscourts.
gov/fedprob/September_2004/juvenile.html
138
Appendix A. Data Tables
Exhibit A-1. Completed interviews by wave, by demographic group and site
Wave 1
NonState
SVORI
SVORI
Adult Males
IA
114
55
IN
64
94
KS
23
48
MD
130
124
ME
35
44
MO
36
50
NV
107
50
OH
47
38
OK
42
51
PA
57
66
SC
179
166
WA
29
48
Total
863
834
Adult Females
IA
35
3
IN
12
101
KS
17
31
ME
7
2
MO
22
0
NV
9
8
OH
15
12
OK
3
7
PA
6
0
SC
24
31
WA
3
9
Total
153
204
Juvenile Males
CO
23
37
FL
40
89
KS
49
20
SC
40
39
Total
152
185
Wave 2
NonSVORI
SVORI
Wave 3
NonSVORI
SVORI
Wave 4
NonSVORI
SVORI
59
49
11
58
20
26
77
25
26
43
123
12
529
29
53
15
63
21
31
31
26
12
50
104
20
455
82
41
14
64
24
27
81
28
29
44
119
12
565
39
56
15
56
26
24
31
27
17
50
95
34
470
87
45
15
65
25
26
82
28
24
46
126
13
582
46
59
24
65
30
35
29
26
27
48
109
33
531
19
10
13
4
18
9
12
3
4
16
2
110
2
62
23
1
0
6
5
5
0
24
6
134
27
12
11
5
16
9
12
2
4
19
2
119
2
68
18
2
0
6
4
3
0
24
7
134
30
11
11
6
19
8
11
1
4
21
2
124
3
75
20
2
0
7
4
4
0
30
7
152
11
37
27
30
105
14
81
10
26
131
9
32
34
33
108
15
74
13
29
131
11
36
28
32
107
18
75
15
33
141
Note: Wave 1 = 30 days pre-release; Wave 2 = 3 months post-release; Wave 3 = 9 month post-release; Wave 4 = 15
months post-release.
A-1
SVORI
Mean
Transition
63.82
Legal
45%
Financial
86%
Public financial
52%
Public health care
76%
Mentor
60%
Documents for employment
55%
Place to live
50%
Transportation
72%
Driver’s license
82%
Access to clothing/food
61%
Health
31.55
Medical treatment
57%
Mental health treatment
23%
Substance use treatment
38%
Victim support group
4%
Anger management
36%
Employment/Education/
Skills
74.40
Job
79%
Education
93%
Money management skills
71%
Life skills
75%
Work on personal
relationships
64%
Change in criminal attitudes
65%
Domestic Violence
6.82
Batterer intervention
8%
Support group
6%
Child
a
Child support payments
Modification in child support
b
debt
a
Modification in custody
a
Parenting skills
a
Child care
Wave 1
NonSVORI
Mean Est.
62.43 1.39
48% 0.14
82% −0.36
54% 0.07
73% −0.14
60% 0.00
56% 0.03
46% −0.16
70% −0.09
81% −0.05
55% −0.26
33.65 −2.10
57% 0.01
28% 0.26
42% 0.16
4% −0.14
37% 0.07
73.60
77%
92%
68%
74%
0.80
−0.16
−0.15
−0.13
−0.04
SE
1.21
0.10
0.14
0.10
0.11
0.10
0.10
0.10
0.11
0.13
0.10
1.21
0.10
0.11
0.10
0.25
0.10
1.25
0.12
0.19
0.11
0.11
OR
1.15
0.70 *
1.07
0.87
1.00
1.03
0.85
0.91
0.95
0.77 *
+
1.02
1.30 *
1.17
0.87
1.07
0.85
0.86
0.88
0.96
63% −0.05 0.10 0.95
69% 0.16 0.11 1.18
8.17 −1.35 1.15
8% 0.00 0.18 1.00
9% 0.40 0.19 1.49 *
SVORI
Mean
48.89
37%
62%
38%
59%
44%
20%
47%
70%
70%
44%
23.55
53%
18%
22%
3%
22%
Wave 2
NonSVORI
Mean
Est.
49.82 −0.92
40% 0.12
61% −0.02
41% 0.12
62% 0.12
39% −0.20
26% 0.34
46% −0.06
70% 0.02
70% 0.01
44% 0.03
25.37 −1.82
56% 0.11
23% 0.33
23% 0.07
2% −0.08
22% 0.00
SE
1.76
0.13
0.13
0.13
0.13
0.13
0.16
0.13
0.14
0.14
0.13
1.52
0.13
0.16
0.16
0.41
0.16
SVORI
Mean
48.52
1.12
43%
0.98
62%
1.13
38%
1.13
58%
0.82
48%
1.40 *
25%
0.94
45%
1.03
62%
1.01
65%
1.03
40%
25.57
1.12
52%
1.39 *
19%
1.08
29%
0.93
3%
1.00
25%
OR
59.73
61%
85%
54%
60%
60.04 −0.31 1.91
63% 0.06 0.13 1.06
85% 0.06 0.18 1.07
57% 0.12 0.13 1.12
58% −0.10 0.13 0.91
63.17
58%
87%
63%
66%
54%
44%
4.35
5%
4%
53% −0.04 0.13 0.96
44% 0.00 0.13 1.00
4.37 −0.03 1.18
6% 0.09 0.28 1.10
3% −0.14 0.36 0.87
Wave 3
NonSVORI
Mean Est.
49.19 −0.67
46% 0.13
65% 0.16
37% −0.05
58% 0.02
46% −0.09
30% 0.25
45% −0.02
64% 0.08
63% −0.07
38% −0.07
26.56 −0.99
53% 0.07
24% 0.26
30% 0.04
2% −0.18
23% −0.09
SE
1.80
0.13
0.13
0.13
0.13
0.13
0.14
0.13
0.13
0.13
0.13
1.55
0.13
0.16
0.14
0.39
0.15
1.20
1.05
0.86
0.89
63.15
59%
88%
62%
64%
58%
48%
3.06
4%
3%
56% −0.08 0.13 0.92
44% −0.16 0.13 0.85
3.47 −0.41 1.01
4% 0.19 0.33 1.21
3% 0.04 0.40 1.04
59%
48%
3.81
4%
3%
41.15 −3.30 2.52
40% 0.29 0.17 1.33 +
40.21
35%
46.07
46%
47.82 −1.75 1.99
47% 0.07 0.13 1.07
38.82
34%
38.99 −0.16 2.46
36% 0.12 0.18 1.13
37.85
33%
87%
35%
60%
39%
86% −0.10 0.36 0.91
37% 0.11 0.13 1.12
63% 0.13 0.13 1.14
39% 0.02 0.13 1.02
87%
30%
53%
29%
86% −0.04 0.44 0.96
31% 0.02 0.18 1.02
49% −0.13 0.17 0.88
31% 0.06 0.18 1.07
79%
31%
53%
28%
62.04 1.12 1.88
62% 0.18 0.13
88% 0.05 0.19
59% −0.15 0.13
64% −0.12 0.13
SVORI
Mean
48.89
1.14
46%
1.17
64%
0.95
36%
1.02
54%
0.91
51%
1.28 +
31%
0.98
45%
1.09
60%
0.93
64%
0.94
37%
27.86
1.07
54%
1.30 +
23%
1.04
32%
0.84
3%
0.92
28%
OR
85% 0.43 0.38
34% 0.17 0.18
52% −0.06 0.17
30% 0.11 0.18
1.53
1.18
0.94
1.12
89%
31%
59%
27%
Wave 4
NonSVORI
Mean Est.
50.04 −1.15
46% 0.03
64% 0.02
39% 0.09
58% 0.17
49% −0.10
34% 0.12
45% 0.01
62% 0.06
65% 0.05
40% 0.11
28.65 −0.79
56% 0.09
27% 0.23
32% 0.02
2% −0.35
26% −0.11
SE
1.80
0.12
0.13
0.13
0.12
0.12
0.13
0.12
0.13
0.13
0.13
1.53
0.12
0.14
0.13
0.40
0.14
OR
1.03
1.02
1.10
1.19
0.90
1.13
1.01
1.06
1.05
1.11
1.09
1.26
1.02
0.70
0.90
64.16 −1.01 1.80
62% 0.15 0.13 1.16
89% 0.10 0.19 1.10
62% 0.02 0.13 1.02
68% 0.19 0.13 1.21
56% −0.15 0.12 0.86
48% 0.02 0.12 1.02
5.76 −1.95 1.22
6% 0.37 0.27 1.45
5% 0.51 0.31 1.67 +
41.44 −1.24 2.28
39% 0.17 0.16 1.18
90% 0.10 0.43
33% 0.06 0.16
53% −0.25 0.16
31% 0.23 0.17
Note: Est. = estimate; SE = standard error; OR = odds ratio. Wave 1 = 30 days pre-release; Wave 2 = 3 months post-release; Wave 3 = 9 month post-release; Wave 4 = 15 months
post-release.
a
Of those with children younger than 18 years of age.
b
Of those who owed back child support.
*p < 0.05.
+p< 0.10.
1.11
1.06
0.78
1.26
Multi-site Evaluation of SVORI: Summary & Synthesis
A-2
Exhibit A-2. Weighted means and parameter estimates of the effect of SVORI on service need (adult males)
Exhibit A-3. Weighted means and parameter estimates of the effect of SVORI on service receipt (adult males)
66%
1.42 0.11
4.14 *
49% 23% 1.14 0.11
57% 25% 1.38 0.11
NA
NA
NA
NA
28.40 16.77 11.63 0.95
75% 52% 1.01 0.11
65% 38% 1.11 0.10
13%
4% 1.25 0.21
14% 11% 0.27 0.15
13%
9% 0.42 0.16
12%
8% 0.39 0.17
41% 26% 0.66 0.11
19%
8% 0.93 0.16
28% 13% 0.94 0.13
19% 12% 0.56 0.14
22%
8% 1.15 0.16
21% 11% 0.74 0.14
35.28 31.21 4.08 1.09
7%
3% 0.91 0.26
33% 26% 0.33 0.11
58% 55% 0.14 0.10
50% 47% 0.11 0.10
16% 19% −0.22 0.13
48% 38% 0.42 0.10
38.16 24.33 13.83 1.39
23%
8% 1.28 0.15
42% 21% 0.97 0.11
3.12 *
4.00 *
NA
*
2.76 *
3.03 *
3.48 *
1.31 +
1.53 *
1.47 *
1.93 *
2.54 *
2.55 *
1.76 *
3.17 *
2.10 *
*
2.48 *
1.40 *
1.15
1.12
0.81 +
1.52 *
*
3.59 *
2.65 *
25%
51%
51%
37%
31%
SE
OR
1.66
*
0.10 2.00 *
0.10 2.89 *
17%
36%
44%
20%
0.48
0.60
0.29
0.87
0.12
0.10
0.10
0.12
1.62 *
1.83 *
1.34 *
2.38 *
Wave 2
NonSVORI SVORI
Mean Mean Est.
57.42 39.56 17.87
44% 18% 1.31
58% 34% 1.02
46%
SE OR
1.90
*
0.15 3.69 *
0.14 2.77 *
22% 1.08 0.15
2.94 *
NA
NA
NA NA
NA
NA
NA
NA NA
NA
82% 85% −0.22 0.17 0.81
13.43 8.87 4.56 0.95
*
NA
NA
NA NA
NA
NA
NA
NA NA
NA
7%
4% 0.58 0.31 1.79 +
20% 16% 0.29 0.17 1.34 +
11%
8% 0.26 0.22 1.30
4%
3% 0.24 0.36 1.28
25% 16% 0.59 0.17 1.80 *
14%
4% 1.32 0.26 3.75 *
12% 11% 0.09 0.21 1.09
15% 12% 0.30 0.19 1.35
12%
7% 0.52 0.23 1.68 *
16%
8% 0.74 0.21 2.10 *
12.75 11.06 1.69 0.95
+
1%
1% 0.35 0.58 1.42
8%
5% 0.52 0.26 1.68 *
25% 23% 0.10 0.15 1.10
5%
7% −0.26 0.27 0.77
9%
8% 0.12 0.24 1.13
28% 23% 0.27 0.15 1.31 +
14.40 8.25 6.15 1.09
*
5%
2% 0.75 0.37 2.12 *
11%
5% 0.76 0.25 2.14 *
7%
18%
11%
34%
4%
10%
8%
20%
0.58
0.64
0.44
0.76
0.28 1.79 *
0.19 1.89 *
0.23 1.55 +
0.15 2.14 *
Wave 3
NonSVORI SVORI
Mean Mean Est.
SE OR
43.17 30.98 12.19 1.97
*
29% 18% 0.59 0.16 1.80 *
42% 24% 0.85 0.14 2.35 *
36%
20%
0.78
0.15 2.19 *
NA
NA
NA
NA NA
NA
NA
NA
NA NA
70% 65% 0.25 0.14 1.28 +
9.23 6.73 2.50 0.81
*
NA
NA
NA
NA NA
NA
NA
NA
NA NA
7%
2% 1.29 0.36 3.64 *
12% 10% 0.20 0.21 1.22
9%
7% 0.20 0.24 1.22
6%
7% −0.25 0.26 0.78
13%
9% 0.43 0.21 1.54 *
13%
6% 0.91 0.24 2.50 *
7%
6% 0.08 0.27 1.09
10%
7% 0.36 0.24 1.44 0
8%
5% 0.58 0.29 1.79 *
7%
7% −0.07 0.25 0.93
12.70 12.50 0.20 0.95
1%
1% −0.55 0.75 0.58
8%
7% 0.14 0.25 1.15
30% 29% 0.04 0.14 1.04
8%
8% −0.09 0.24 0.92
8%
7% 0.14 0.25 1.15
22% 23% −0.04 0.16 0.96
11.48 7.75 3.74 1.05
*
4%
2% 0.80 0.42 2.23 +
8%
5% 0.66 0.28 1.94 *
9%
16%
11%
21%
6%
10%
10%
14%
0.44 0.26 1.55 +
0.48 0.20 1.62 *
0.12 0.21 1.13
0.49 0.18 1.63 *
Wave 4
NonSVORI SVORI
Mean Mean Est.
29.76 24.41 5.35
24% 15% 0.55
31% 22% 0.43
27%
16%
SE OR
1.79
*
0.17 1.73 *
0.15 1.54 *
0.64 0.17 1.89 *
NA
NA
NA NA NA
NA
NA
NA NA NA
53% 53% 0.00 0.12 1.00
6.00 6.01 −0.01 0.76
NA
NA
NA NA NA
NA
NA
NA NA NA
5%
2% 0.69 0.38 2.00 +
7% 11% −0.47 0.24 0.62 *
6%
7% −0.16 0.26 0.86
6%
8% −0.31 0.27 0.73
7%
6% 0.09 0.27 1.10
9%
4% 0.79 0.28 2.20 *
4%
5% −0.28 0.32 0.76
6%
5% 0.10 0.30 1.11
5%
5% 0.12 0.31 1.13
5%
6% −0.18 0.29 0.84
10.86 11.93 −1.07 0.96
2%
0% 2.08 1.06 7.99 +
3%
5% −0.41 0.35 0.66
27% 27% −0.02 0.15 0.98
9% 10% −0.09 0.22 0.91
7%
9% −0.31 0.25 0.73
18% 21% −0.19 0.17 0.83
8.47 5.88 2.59 0.93
*
3%
2% 0.60 0.47 1.83
6%
3% 0.72 0.34 2.06 *
7%
10%
10%
14%
6%
8%
7%
10%
0.25 0.27 1.28
0.24 0.24 1.27
0.41 0.25 1.51 +
0.44 0.22 1.55 *
(continued)
Appendix A — Data Tables
A-3
Coordination
Needs assessment
Case manager
Worked with someone to
reintegrate
Needs assess specific for
release
Reentry plan developed
Currently on probation/parole
Transition
Programs prepare for release
Class specifically for release
Financial assistance
Public financial assistance
Public health care
Legal assistance
Documents for employment
Mentoring
Place to live
Transportation
Driver’s license
Access to clothing/food
Health
Victim support group
Anger management program
Medical treatment
Dental services
Mental health treatment
Substance use treatment
Employment/education/ skills
Money management skills
Life skills
Work on personal
relationships
Change criminal attitudes
Any educational services
Any employment services
Wave 1
NonSVORI SVORI
Mean Mean Est.
59.52 33.00 26.52
63% 46% 0.69
66% 41% 1.06
Domestic violence
Support group
Batterer intervention
Child
a
Child care
Modification in child support
b
debt
a
Child support payments
a
Modification in custody
a
Parenting classes
Wave 1
NonSVORI SVORI
Mean Mean Est.
7.43 4.72 2.71
10%
6% 0.60
5%
4% 0.28
11.26 5.74 5.53
8%
3% 1.01
22%
6%
4%
25%
11%
2%
2%
15%
0.84
1.22
0.54
0.63
SE
OR
1.00
*
0.19 1.82 *
0.25 1.33
0.99
*
0.32 2.75 *
0.33
0.37
0.39
0.16
2.32 *
3.39 *
1.72
1.88 *
Wave 2
NonSVORI SVORI
Mean Mean Est.
0.97 1.17 −0.21
1%
2% −0.35
1%
1% 0.04
4.19 1.37 2.83
4%
0% 2.34
14%
3%
3%
5%
SE
0.60
0.56
0.68
0.74
1.04
OR
0.70
1.04
*
10.42 *
4% 1.49 0.58 4.43 *
2% 0.49 0.62 1.62
0%
.
.
.
3% 0.54 0.45 1.72
Wave 3
NonSVORI SVORI
Mean Mean Est.
SE
1.35 1.40 −0.05 0.64
2%
2% 0.11 0.47
1%
1% −0.40 0.77
2.75 2.14 0.61 0.69
2%
2% 0.36 0.58
11%
2%
1%
4%
4% 1.07
1% 0.47
3% −0.86
3% 0.36
0.61
0.87
0.64
0.52
OR
1.11
0.67
1.44
2.92 +
1.60
0.42
1.44
Wave 4
NonSVORI SVORI
Mean Mean Est.
1.02 1.64 −0.62
1%
2% −0.57
1%
1% −0.31
1.71 1.68 0.03
2%
1% 0.72
5%
1%
2%
1%
SE
0.67
0.53
0.74
0.58
0.72
Of those with children younger than 18 years of age.
b
Of those who owed back child support.
*p < 0.05.
+p < 0.10.
0.57
0.73
2.05
5% −0.05 0.60 0.95
1% −0.30 1.01 0.74
1% 0.35 0.74 1.42
2% −0.70 0.64 0.50
Note: NA = nonapplicable; Est. = estimate; SE = standard error; OR = odds ratio. Wave 1 = 30 days pre-release; Wave 2 = 3 months post-release; Wave 3 = 9 month post-release;
Wave 4 = 15 months post-release.
a
OR
Multi-site Evaluation of SVORI: Summary & Synthesis
A-4
Exhibit A-3. Weighted means and parameter estimates of the effect of SVORI on service receipt (adult males) (continued)
Exhibit A-4. Service needs weighted means and SVORI parameter estimates (adult females)
78.76 1.73
81% −0.52
95%
70% −0.15
75% 0.10
SE
2.47
23%
34%
24%
40%
30%
23%
23%
27%
28%
26%
2.94
27%
23%
24%
24%
23%
SVORI
Mean
59.92
1.02
49%
0.58
75%
1.70 * 51%
0.82
80%
0.78
56%
1.11
29%
0.79
57%
0.51 * 75%
0.74
71%
0.88
56%
38.52
0.92
71%
0.96
43%
0.82
35%
1.11
19%
1.96 * 25%
OR
2.65
31% 0.59
Wave 2
NonSVORI
Mean Est.
59.63 0.29
46% 0.14
71% 0.21
61% −0.42
79% 0.06
64% −0.31
32% −0.14
52% 0.20
66% 0.42
67% 0.22
58% −0.08
40.72 −2.19
70% 0.02
38% 0.22
38% −0.12
20% −0.09
38% −0.59
SE
3.43
0.27
0.30
0.27
0.34
0.27
0.29
0.27
0.30
0.29
0.27
4.05
0.30
0.28
0.28
0.34
0.30
OR
1.15
1.23
0.66
1.07
0.73
0.87
1.22
1.52
1.24
0.92
1.03
1.25
0.89
0.91
0.55 *
64.01 −0.37
65% 0.22
92%
55% −0.06
64% −0.17
3.83
0.29 1.24
25% 0.86
25% 1.10
63.64
70%
93%
54%
60%
79% 0.04
74% −0.19
20.97 0.44
16% −0.24
26% 0.14
47.88 4.02
44% −0.20
27%
26%
3.92
30%
26%
3.49
26%
59%
46%
8.04
7%
9%
40.80
46%
62%
47%
8.12
4%
12%
46.06
42%
−0.11
−0.02
−0.09
0.28 0.90
0.27 0.98
3.25
−0.32
−5.25
0.16
91%
41% −0.04
71% 0.09
35% −0.41
26% 0.96
28% 1.09
27% 0.66
96%
34%
50%
28%
100%
46% −0.51
68% −0.78
24% 0.18
1.04
0.83
0.79
1.15
0.82
SVORI
Mean
51.11
38%
71%
44%
67%
57%
28%
43%
64%
56%
43%
37.37
61%
40%
36%
18%
32%
Wave 3
NonSVORI
Mean Est.
52.04 −0.93
37% 0.06
66% 0.22
58% −0.57
69% −0.08
58% −0.01
30% −0.11
40% 0.09
57% 0.30
60% −0.15
46% −0.11
38.27 −0.90
56% 0.22
36% 0.18
41% −0.20
18% −0.02
41% −0.38
SE
3.67
0.28
0.28
0.27
0.29
0.27
0.30
0.27
0.27
0.27
0.27
3.85
0.27
0.28
0.28
0.35
0.28
OR
1.06
1.25
0.56
0.92
0.99
0.90
1.09
1.35
0.86
0.89
1.25
1.20
0.82
0.98
0.68
60.24
57%
87%
54%
51%
60.36
57%
87%
61%
58%
−0.12
0.01
−0.01
−0.27
−0.26
4.02
0.27
0.41
0.27
0.27
0.44 0.73
4.46
0.32 1.17
67%
45%
11.50
8%
15%
38.45
41%
59% 0.35
40% 0.20
11.85 −0.35
8% 0.04
16% −0.07
39.45 −1.01
41% 0.02
0.27
0.27
3.74
0.48
0.38
5.11
0.32
0.33 0.60
0.33 0.46 *
0.37 1.20
96%
27%
57%
23%
100%
32% −0.20
57% 0.00
27% −0.21
0.35 0.82
0.31 1.00
0.37 0.81
0.27 0.94
0.28 0.84
1.01
0.99
0.76
0.77
1.41
1.22
1.04
0.93
1.02
Wave 4
NonSVORI SVORI
Mean Mean Est.
47.67 56.33 −8.66
36%
43% −0.30
63%
65% −0.08
* 43%
64% −0.84
63%
70% −0.31
50%
60% −0.42
26%
37% −0.52
51%
48%
0.11
54%
64% −0.42
52%
61% −0.36
40%
52% −0.48
35.25 39.35 −4.10
58%
57%
0.02
38%
41% −0.11
33%
36% −0.13
20%
18%
0.12
27%
44% −0.74
SE
3.88
0.26
0.26
0.26
0.27
0.26
0.28
0.25
0.26
0.26
0.26
3.90
0.26
0.26
0.26
0.32
0.27
3.95
0.26
0.41
0.25
0.26
Regression results not shown when cell sizes <10.
Of those with children younger than 18 years of age.
b
Of those who owed back child support.
*p < 0.05 for test of significant difference between SVORI and non-SVORI.
*
0.74
0.92
0.43 *
0.74
0.66
0.60
1.12
0.66
0.70
0.62
1.02
0.90
0.88
1.13
0.48 *
57.49
56%
87%
46%
50%
63.68
63%
92%
59%
63%
−6.20
−0.28
−0.50
−0.50
−0.53
64%
43%
9.57
6%
14%
34.54
37%
61%
45%
10.97
8%
14%
42.90
40%
0.11 0.26 1.11
−0.10 0.26 0.90
−1.40 3.17
70%
25%
52%
20%
90%
35%
63%
31%
0.75
0.61
0.61
0.59 *
0.00 0.37 1.00
−8.36 4.38
−0.09 0.30 0.91
−0.51 0.32 0.60
−0.44 0.29 0.65
−0.57 0.34 0.57
Note: Est. = estimate; SE = standard error; OR = odds ratio. Wave 1 = 30 days pre-release; Wave 2 = 3 months post-release; Wave 3 = 9 month post-release; Wave 4 = 15 months
post-release.
a
OR
A-5
Appendix A — Data Tables
SVORI
Mean
Transition
73.97
Legal
54%
Financial
91%
Public financial
60%
Public health care
91%
Mentor
85%
Documents for employment 58%
Place to live
58%
Transportation
83%
Driver’s license
82%
Clothing/food banks
78%
Health
54.94
Medical treatment
80%
Mental health treatment
56%
Substance use treatment
68%
Victim support group
30%
Anger management
42%
Employment/Education/
skills
80.48
Job
88%
Education
95%
Money management skills
73%
Life skills
73%
Work on personal
relationships
78%
Change in criminal attitudes 78%
Domestic Violence
21.41
Batterer intervention
19%
Support group
24%
Child
51.91
a
Child support payments
49%
Modification in child support
b
debt
85%
a
Modification in custody
42%
a
Parenting skills
70%
a
Child care
45%
Wave 1
NonSVORI
Mean Est.
71.82 2.15
55% 0.02
85% −0.55
72% 0.53
90% −0.20
82% −0.24
60% 0.10
52% −0.24
72% −0.67
77% −0.30
76% −0.12
57.42 −2.48
78% −0.09
55% −0.04
64% −0.20
32% 0.11
59% 0.67
Coordination
Needs assessment
Case manager
Worked with someone to
reintegrate
Currently on probation/parole
Transition
Program to prepare for
release
Class to prepare for release
Financial assistance
Public financial assistance
Public health care
Legal assistance
Documents for employment
Mentoring
Place to live
Transportation
Driver’s license
Access to clothing/food
Health
Victim support group
Anger management program
Medical treatment
Dental services
Mental health treatment
Substance use treatment
Employment/Education/Skills
Money management skills
Life skills
Work on personal
relationships
Change criminal attitudes
Any educational services
Any employment services
Wave 1
NonSVORI SVORI
Mean Mean Est.
72.67 29.11 43.56
79%
46%
1.49
70%
33%
1.58
SE
OR
3.44
*
0.27 4.44 *
0.24 4.83 *
81%
25%
2.53 0.28 12.62 *
40.13
23.02 17.11 2.20
Wave 2
NonSVORI SVORI
Mean Mean Est. SE OR
69.35 36.50 32.85 3.82
*
52%
15%
1.78 0.31 5.92 *
74%
38%
1.53 0.30 4.61 *
Wave 3
NonSVORI SVORI
Mean Mean Est.
55.05 29.77 25.28
39%
22% 0.80
59%
27% 1.34
*
68%
88%
26.33
18%
2.25 0.31 9.50 *
75%
0.89 0.37 2.44 *
12.45 13.87 2.31
*
51%
80%
16.60
16%
56%
10.19
89%
82%
19%
24%
25%
22%
59%
36%
37%
23%
25%
43%
47.67
17%
37%
68%
63%
49%
52%
53.64
38%
63%
58%
1.75 0.32 5.73 *
47%
1.63 0.27 5.10 *
3%
13%
0.72 0.30 2.05 *
9%
1.13 0.34 3.10 *
14%
0.52 0.29 1.68
43%
0.61 0.23 1.85 *
10%
1.59 0.30 4.89 *
18%
0.99 0.26 2.68 *
17%
0.38 0.27 1.46
15%
0.63 0.28 1.89 *
29%
0.62 0.23 1.86 *
35.59 12.08 2.68
*
8%
0.82 0.34 2.27 *
18%
0.96 0.26 2.60 *
71% −0.15 0.25 0.86
45%
0.74 0.23 2.11 *
34%
0.60 0.23 1.82 *
37%
0.63 0.23 1.88 *
24.66 28.97 3.12
*
10%
1.74 0.29 5.68 *
25%
1.62 0.24 5.04 *
12%
39%
30%
4%
29%
35%
29%
37%
8%
41%
24.66
8%
4%
39%
18%
32%
48%
25.69
11%
21%
2%
30%
0.40 0.29 1.49
25%
0.22 0.30 1.25
4%
12%
1.12 0.34 3.05 *
7%
11%
1.15 0.37 3.16 *
12%
1.43 0.35 4.17 *
6%
16%
1.32 0.31 3.74 *
14.17 10.49 2.36
*
0%
3%
44% −0.19 0.27 0.83
3%
17%
0.83 0.32 2.28 *
19%
1.36 0.30 3.88 *
7.39 18.30 2.87
*
2%
6%
10%
25%
25%
7%
18%
22%
13%
22%
7%
17%
25.70
7%
5%
54%
20%
26%
42%
16.90
10%
14%
1%
20% 0.27
26% −0.05
10%
10% 0.74
6%
7% 0.69
7% 1.28
3%
12% 0.39
17.78 7.92
0%
7%
50% 0.17
15% 0.32
14% 0.80
20% 1.03
7.58 9.32
3%
5%
46%
64%
61%
50%
19%
29%
40%
25%
21%
33%
21%
48%
7%
7%
6%
16%
13%
24%
18%
23%
8%
10%
9%
10%
1.27
1.50
0.86
1.09
0.25
0.24
0.23
0.24
3.57
4.48
2.36
2.97
*
*
*
*
1.87 0.40 6.47 *
1.60 0.31 4.93 *
1.73
1.15
6.41
0.57
0.97
0.79
0.97
SE OR
4.25
*
0.29 2.23 *
0.28 3.83 *
0.31 5.64 *
0.31 3.14 *
2.14
*
0.33 1.31
0.31 0.95
0.41 2.10
0.44 1.99
0.42 3.60 *
0.37 1.48
2.47
*
0.27
0.36
0.32
0.30
2.57
0.42
0.37
0.41
0.36
1.19
1.37
2.24 *
2.79 *
*
1.76
2.65 *
2.21
2.65 *
Wave 4
NonSVORI SVORI
Mean Mean Est. SE OR
39.27 25.47 13.80 4.06
*
36%
21% 0.74 0.30 2.10 *
42%
25% 0.77 0.28 2.17 *
39%
54%
13.16
18%
43%
11.68
11%
20%
26%
12%
10%
17%
8%
12%
4%
12%
21.69
2%
4%
53%
18%
20%
33%
14.14
6%
10%
4%
24%
27%
8%
10%
7%
8%
8%
4%
18%
18.90
2%
5%
54%
18%
18%
17%
7.80
5%
7%
7%
21%
18%
22%
6%
9%
11%
11%
1.06 0.30 2.90 *
0.42 0.26 1.53
1.48 2.02
−0.18
−0.02
0.36
−0.04
1.04
0.33
0.31
0.44
0.44
0.43
0.83
0.98
1.44
0.96
2.83 *
0.50 0.45 1.65
−0.49 0.38 0.61
2.79 2.37
−0.03
0.01
0.17
0.90
6.34
0.27
0.35
0.33
0.32
2.25
0.97
1.01
1.19
2.47 *
*
0.51 0.46 1.66
1.04 0.38 2.83 *
0.60 0.38 1.83
0.87 0.36 2.39 *
(continued)
Multi-site Evaluation of SVORI: Summary & Synthesis
A-6
Exhibit A-5. Weighted means and parameter estimates of the effect of SVORI on service receipt
Exhibit A-5. Weighted means and parameter estimates of the effect of SVORI on service receipt (continued)
Domestic Violence
Support group
Batterer intervention
Child
a
Child care
Modification in child support
b
debt
a
Child support payments
a
Modification in custody
a
Parenting classes
SVORI
Mean
15.36
23%
7%
19.39
10%
23%
3%
13%
51%
Wave 1
NonSVORI
Mean Est.
8.46
6.91
12%
0.85
5%
0.34
9.74 9.66
2%
0%
1%
9%
28%
SE
3.03
0.30
0.45
2.54
0.49 0.43
1.00 0.28
OR
*
2.33 *
1.40
*
1.63
2.71 *
Wave 2
NonSVORI SVORI
Mean Mean Est. SE
3.11
0.64 2.46 1.25
5%
1%
1%
0%
5.68
5.40 0.29 2.09
5%
4%
20%
7%
1%
7%
8%
4%
7%
7%
OR
*
Wave 3
NonSVORI SVORI
Mean Mean Est.
SE
4.23
1.20 3.03 2.03
5%
2%
3%
1%
4.98
1.36 3.62 1.68
5%
1%
6%
7%
4%
5%
0%
3%
1%
1%
OR
*
Wave 4
NonSVORI SVORI
Mean Mean Est. SE
5.96
1.67 4.29 2.24
8%
1%
4%
2%
6.32
5.17 1.15 1.88
7%
5%
12%
7%
2%
8%
OR
8%
5%
2%
8%
Note: Est. = estimate; SE = standard error; OR = odds ratio. Regression results not shown when cell sizes <10. Wave 1 = 30 days pre-release; Wave 2 = 3 months post-release; Wave
3 = 9 month post-release; Wave 4 = 15 months post-release.
a
Of those with children younger than 18 years of age.
b
Of those who owed back child support.
Appendix A — Data Tables
A-7
Transition
Legal assistance
After-school/weekend/
a
summer sports program
Financial assistance
Public financial assistance
Public health care
Mentor
Documents for employment
Place to live
Transportation
Driver’s license
Clothing/food banks
Health
Medical treatment
Mental health treatment
Substance use treatment
Victim support group
Anger management
Employment/Education/
Skills
Job
Education
Money management skills
Life skills
Work on personal
relationships
Change in criminal attitudes
Domestic Violence
Batterer intervention
Support group
Wave 1
NonSVORI SVORI
Mean Mean Est.
52
49
2.57
63%
51%
0.50
Wave 2
NonSVORI SVORI
SE OR
SE
OR
Mean Mean Est.
2.70
38
39
−0.48
3.20
0.25 1.65 *
33%
37% −0.20 31.00 0.82
NA
50%
21%
50%
65%
59%
24%
66%
90%
41%
30
44%
25%
29%
5%
48%
NA
48%
24%
55%
59%
52%
28%
62%
91%
35%
32
47%
19%
36%
3%
57%
0.08
−0.21
−0.17
0.25
0.27
−0.22
0.17
−0.09
0.26
−2.04
−0.13
0.34
−0.28
0.59
−0.36
0.24
0.29
0.24
0.25
0.24
0.27
0.25
0.40
0.25
3.17
0.24
0.29
0.26
0.59
0.24
75
86%
95%
64%
78%
75
0.66
88% −0.19
95% −0.05
66% −0.09
71%
0.33
2.74
0.38
0.48
0.25
0.26
60%
70%
7
6%
9%
53%
0.27
75% −0.26
9
−1.30
8%
10% −0.12
0.24 1.31
0.26 0.77
3.00
0.88
1.40
0.75
1.80
0.70
33%
25%
13%
40%
38%
27%
28%
65%
86%
33%
18
37%
9%
13%
0%
29%
26%
35%
17%
39%
32%
18%
38%
72%
82%
34%
15
33%
9%
10%
2%
22%
0.31
−0.48
−0.34
0.05
0.28
0.49
−0.41
−0.31
0.29
−0.07
2.30
0.18
−0.02
0.30
0.33
0.83
0.96
0.92
1.38
56
74%
88%
46%
52%
52
69%
87%
44%
42%
40%
34%
9
7%
12%
36%
34%
4
4%
5%
1.08
0.81
0.84
1.29
1.31
0.80
1.19
0.92
1.29
0.43 0.89
0.34
0.31
0.40
0.29
0.30
0.35
0.30
0.31
0.37
0.30
3.02
0.29
0.49
0.44
1.37
0.62
0.71
1.05
1.33
1.64
0.66
0.73
1.34
0.93
Wave 3
NonSVORI SVORI
OR
Mean Mean Est. SE
40
41
−0.86 3.47
49%
37% 0.50 0.30 1.60
25%
35%
19%
42%
31%
30%
49%
66%
75%
39%
15
27%
11%
11%
1%
24%
0.19
−0.42
−0.71
−0.15
0.12
0.23
−0.35
−0.24
−0.27
0.15
2.18
0.25
−0.04
0.06
0.33 1.40
28%
27%
10%
39%
34%
35%
40%
60%
69%
43%
17
32%
11%
11%
0%
30%
3.56
0.21
0.05
0.06
0.42
4.19
0.33
0.41
0.29
0.29
63
72%
91%
52%
72%
0.17
0.01
5.15
0.30 1.19
0.30 1.01
3.97
46%
47%
6
6%
7%
1.19
0.98
1.35
1.24
1.06
1.06
1.52
0.38
0.31
0.39
0.29
0.30
0.32
0.29
0.31
0.34
0.30
2.98
0.32
0.44
0.48
0.33 0.33 1.39
20%
47%
20%
40%
37%
21%
50%
74%
76%
41%
18
38%
16%
10%
2%
24%
54
66%
85%
46%
55%
9.54
0.27
0.62
0.23
0.76
58
65%
85%
49%
64%
59
−1.57
74% −0.42
91% −0.50
42%
0.28
62%
0.08
33%
38%
3
4%
2%
0.55 0.30 1.73
0.35 0.29 1.42
2.80 3.39
44%
38%
3
3%
2%
40%
0.16 0.29 1.17
46% −0.33 0.30 0.72
5
−2.50 2.77
5%
5%
(continued)
1.29
0.96
1.06
*
1.31
1.86
1.26
2.14 *
0.20
−0.47
−0.16
−0.28
−0.30
0.55
−0.54
−0.66
−0.44
−0.51
−0.56
−0.13
SE OR
3.43
0.30 0.83
23%
36%
17%
34%
30%
31%
37%
59%
67%
29%
17
35%
9%
11%
2%
29%
4.00
0.31
0.45
0.29
0.29
1.21
0.66
0.49
0.86
1.13
1.26
0.70
0.79
0.76
1.16
Wave 4
NonSVORI SVORI
Mean Mean Est.
37
42
−5.74
31%
35% −0.18
0.45
0.29
0.36
0.29
0.30
0.32
0.28
0.31
0.30
0.30
2.78
0.30
1.22
0.63
0.85
0.76
0.74
1.73
0.58
0.52 *
0.65
0.60
0.88
0.21
0.41 1.23
0.28
0.31 1.32
3.99
0.30
0.40
0.29
0.28
0.66
0.60
1.32
1.09
Multi-site Evaluation of SVORI: Summary & Synthesis
A-8
Exhibit A-6. Service needs weighted means and SVORI parameter estimates (juvenile males)
Exhibit A-6. Service needs weighted means and SVORI parameter estimates (juvenile males) (continued)
Child
b
Child support payments
Modification in child support
b
debt
c
Modification in custody
a
Parenting skills
a
Child care
All Services Need
Wave 1
NonSVORI SVORI
Mean Mean Est.
SE OR
29
27
2.91 11.71
22%
16%
0%
19%
64%
17%
49
100%
24%
42%
24%
49
0.61
2.18
Wave 2
NonSVORI SVORI
Mean Mean Est.
SE
19
41
−21.69 12.20
5%
27%
0%
16%
39%
17%
36
100%
41%
60%
35%
35
1.32
2.75
OR
Wave 3
NonSVORI SVORI
Mean Mean Est.
50
26
17%
11%
100%
28%
80%
72%
38
100%
17%
42%
32%
35
SE
2.81 2.69
OR
Wave 4
NonSVORI SVORI
Mean Mean Est.
24
35 −10.22
4%
29%
0%
19%
56%
18%
35
100%
16%
59% −0.10
33%
38
−3.60
SE OR
8.87
0.62 0.90
2.69
Notes: Est. = estimate; SE = standard error; OR = odds ratio. Wave 1 = 30 days pre-release; Wave 2 = 3 months post-release; Wave 3 = 9 month post-release; Wave 4 = 15 months
post-release.
Regression results not shown when cell sizes <10. NS = non-SVORI. NA = not applicable.
a
Of those who were not reincarcerated at the time of the interview.
b
Of those with children.
c
Of those who owed back child support.
*p < 0.05.
Appendix A — Data Tables
A-9
Wave 1
Wave 2
NonNonSVORI SVORI
SVORI SVORI
SE OR
Mean Mean Est.
Mean Mean Est.
Coordination
57
53
4.90 2.89
43
42
1.64
Needs assessment
83% 79%
0.22 0.30 1.25
74% 54%
0.89
Meeting with case manager 89% 88%
0.18 0.38 1.19
66% 66%
−0.01
Work with someone to
reintegrate
76% 66%
0.49 0.29 1.63
40% 39%
0.03
Assistance accessing child
welfare caseworker
23% 25%
−0.09 0.28 0.91
9%
9%
−0.06
Meeting with child welfare
caseworker
41% 33%
0.36 0.26 1.44
17% 16%
0.07
Current probation/parole
NA
NA
90% 82%
0.65
Transition
23
21
2.10 2.10
12
8
3.49
Taken programs to prepare
63% 52%
0.41 0.25 1.51
NA
NA
for release
Taken class specifically for
release
42% 42%
0
0.24 1.00
NA
NA
After-school/weekend/
a
summer sports program
NA
NA
9% 18%
Financial assistance
8%
6%
0.31 0.43 1.36
13%
7%
0.74
Public financial assistance
3%
2%
2%
2%
Public health care
9% 17%
−0.75 0.35 0.47 *
16%
8%
0.77
Legal assistance
23% 21%
0.10 0.30 1.11
17%
7%
1.08
Documents for employment 20% 18%
0.13 0.31 1.14
18% 11%
0.63
Mentoring
37% 42%
−0.21 0.25 0.81
9% 15%
Place to live
31% 18%
0.68 0.28 1.97 *
10%
6%
Transportation
21% 17%
0.31 0.29 1.36
18% 10%
0.70
Driver’s license
16% 16%
0.00 0.32 1.00
12% 10%
0.27
Access to clothing/food
9%
6%
0.48 0.44 1.61
9%
7%
Health
42
47
−5.40 2.70
*
12
7
4.62
Victim support group
6%
6%
0%
0%
Anger management
48% 63%
−0.61 0.24 0.55 *
15%
8%
0.62
program
Medical treatment
69% 69%
−0.02 0.27 0.98
20% 15%
0.34
Dental services
42% 54%
−0.46 0.24 0.63
14%
8%
0.56
Mental health treatment
25% 31%
−0.29 0.27 0.75
9%
5%
Substance use treatment
56% 58%
−0.06 0.25 0.95
16%
8%
0.80
Wave 3
NonSVORI SVORI
OR
Mean Mean Est. SE
24
18
6.25 3.07
*
23%
12%
0.82 0.40 2.27 *
39%
30%
0.42 0.31 1.52
SE
3.79
0.31
0.30
OR
0.30
1.03
29%
16%
0.57
0.94
6%
3%
0.44
0.49
1.81
1.07
1.92
7%
45%
8
7%
45%
7
2.42 *
0.99
0.53
2.10
0.42
0.49
0.39
2.17
2.94 *
1.88
0.42
0.48
2.02
1.31
2.23
*
0.48
0.36
0.47
1.85
1.40
1.75
0.42
2.22
0.76
0.00
0.41
0.35
2.13 *
0.29 1.00
1.58
Wave 4
NonSVORI SVORI
Mean Mean Est.
13
14
−1.30
9%
16%
14%
27% −0.83
18%
13%
2%
3%
1%
39%
8
5%
28%
5
NA
NA
NA
NA
NA
NA
NA
NA
8%
6%
1%
14%
14%
15%
11%
2%
8%
6%
4%
8
3%
16%
1%
3%
8%
4%
16%
9%
5%
14%
8%
3%
10
1%
6%
2%
1%
6%
11%
11%
4%
7%
19%
17%
5%
10
0%
16%
1%
4%
5%
6%
4%
12%
2%
7%
6%
1%
7
0%
6%
14%
12%
7%
8%
7%
21%
13%
5%
10%
5%
26%
14%
6%
8%
5%
12%
10%
2%
10%
0.65
−0.14
0.24
0.46
1.91
0.40 0.87
0.48 1.27
−1.34
2.04
−0.55
−0.15
0.38
0.43
0.57
0.86
SE
2.61
OR
0.37 0.44 *
0.37
0.44 1.44
0.49
2.99
0.31 1.63
1.45
*
1.18
0.45 3.24 *
3.19
1.93
0.93
0.46
0.37 2.55 *
0.47 1.58
(continued)
Multi-site Evaluation of SVORI: Summary & Synthesis
A-10
Exhibit A-7. Service receipt weighted means and SVORI parameter (juvenile males)
Exhibit A-7. Service receipt weighted means and SVORI parameter (juvenile males) (continued)
Wave 1
NonSVORI SVORI
Mean Mean Est.
Employment/Education/
Skills
Money management skills
Life skills
Work on personal
relationships
Change in criminal attitudes
Any educational services
Any employment services
Domestic Violence
Support group
Batterer intervention
Child
b
Child support payments
Modification in child support
b
debt
c
Modification in custody
a
Parenting skills
a
Child care
All Services Receipt
52
22%
50%
50
18%
47%
38%
68%
93%
42%
7
11%
3%
17
0%
35%
80%
96%
27%
5
9%
2%
6
0%
0%
6%
50%
13%
38
0%
0%
12%
12%
37
SE OR
2.04 3.02
0.27 0.29 1.31
0.11 0.24 1.12
0.12
−0.61
−0.55
0.69
1.48
0.23
0.25
0.28
0.50
0.25
2.20
0.42
10.97 6.48
1.23 1.85
1.13
0.54 *
0.58
2.00 *
1.26
Wave 2
NonSVORI SVORI
Mean Mean Est.
21
7%
6%
16
5%
11%
5%
22%
56%
29%
0
0%
0%
0
0%
11%
24%
28%
15%
1
2%
0%
2
0%
0%
0%
0%
0%
19
0%
0%
0%
10%
16
SE
5.02 2.38
−0.10
1.19
0.81
−0.81
0.36
0.30
0.36
0.60
Wave 3
NonSVORI SVORI
Mean Mean Est.
OR
*
0.91
3.28 *
2.25 *
−1.93 1.91
3.33 1.61
*
18
4%
13%
13
5%
9%
11%
28%
30%
21%
0
0%
1%
4
0%
6%
16%
29%
15%
0
0%
0%
2
0%
0%
0%
14%
0%
12
100%
0%
0%
0%
10
SE
4.29
2.65
0.44
0.44
OR
1.55
0.67 0.38 1.95
0.08 0.32 1.09
0.37 0.37 1.45
2.05
1.52
Wave 4
NonSVORI SVORI
Mean Mean Est.
14
3%
10%
9
3%
7%
5%
16%
35%
16%
0
0%
0%
3
0%
2%
12%
22%
10%
0
1%
0%
1
0%
0%
4%
4%
4%
9
0%
0%
5%
0%
7
SE
OR
4.65
2.49
0.97
0.61
0.54
0.74 2.65
0.36 1.83
0.44 1.71
1.69
3.29
2.02
1.24
Notes: Est. = estimate; SE = standard error; OR = odds ratio. Wave 1 = 30 days pre-release; Wave 2 = 3 months post-release; Wave 3 = 9 month post-release; Wave 4 = 15 months
post-release.
Regression results not shown when cell sizes <10. NS = non-SVORI. NA = not applicable.
a
Asked only of respondents who were not reincarcerated at the time of the interview.
b
Asked only of respondents with children.
c
Asked only of respondents who owed back child support.
A-11
Appendix A — Data Tables
*p < 0.05